Ò
PAIN 155 (2014) 2551–2559
www.elsevier.com/locate/pain
Tactile allodynia in patients with lumbar radicular pain (sciatica) Ruth Defrin a, Marshall Devor b,⇑,1, Silviu Brill c,1 a
Department of Physical Therapy, School of Allied Health Professions, The Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel Department of Cell and Developmental Biology, Institute of Life Sciences, and Center for Research on Pain, The Hebrew University of Jerusalem, Jerusalem 91904, Israel c Institute of Pain Medicine, Department of Anesthesia and Critical Care Medicine, Tel Aviv Sourasky Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel-Aviv 69978, Israel b
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
a r t i c l e
i n f o
Article history: Received 25 March 2014 Received in revised form 28 August 2014 Accepted 11 September 2014
Keywords: Central sensitization Radicular pain Sciatica Shooting pain Tactile allodynia
a b s t r a c t We report a novel symptom in many patients with low back pain (LBP) that sheds new light on the underlying pain mechanism. By means of quantitative sensory testing, we compared patients with radicular LBP (sciatica), axial LBP (LBP without radiation into the leg), and healthy controls, searching for cutaneous allodynia in response to weak tactile and cooling stimuli on the leg and low back. Most patients with radicular pain (~60%) reported static and dynamic tactile allodynia, as well as cooling allodynia, on the leg, often extending into the foot. Some also reported allodynia on the low back. In axial LBP, allodynia was almost exclusively on the back. The degree of dynamic tactile allodynia correlated with the degree of background pain. The presence of allodynia suggests that the peripheral nerve generators of background leg and back pain have also induced central sensitization. The distal (foot) location of the allodynia in patients who have it indicates that the nociceptive drive that maintains the central sensitization arises paraspinally (ectopically) in injured ventral ramus afferents; this is not an instance of somatic referred pain. The presence of central sensitization also provides the first cogent account of shooting pain in sciatica as a wave of activity sweeping vectorially across the width of the sensitized dorsal horn. Finally, the results endorse leg allodynia as a pain biomarker in animal research on LBP, which is commonly used but has not been previously validated. In addition to informing the underlying mechanism of LBP, bedside mapping of allodynia might have practical implications for prognosis and treatment. Social media question: How can you tell whether pain radiating into the leg in a patient with sciatica is neuropathic, ie, due to nerve injury? Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
1. Introduction Low back pain (LBP) is consistently ranked among the leading causes of disability due to disease or injury [44]. Treatment is often unsatisfactory, at least in part as a result of uncertainty concerning the underlying mechanisms. In particular, radiological evidence of spinal pathology correlates imperfectly with pain symptoms [4,12,35] suggesting that it is not the pathology per se that triggers pain, but rather secondary changes in neural function that are triggered by the pathology in susceptible individuals. Here we report new observations on symptomatology that shed light on LBP mechanisms and that may have direct implications for diagnosis, prognosis, and treatment. ⇑ Corresponding author. E-mail address: [email protected] (M. Devor). The last 2 authors contributed equally to this article, and both should be considered senior author. 1
LBP, which frequently also features pain in the leg (lumbar radicular pain, or sciatica), is variously considered to reflect the activation of nociceptive sensory endings (normal or sensitized by inflammation), or a neuropathic process in the peripheral nervous system [1,4,8,10,12,13,22,31]. According to the nociception/ inflammation hypothesis, pain results from electrical impulses originating in sensory endings in deep peripheral tissues (eg, disc, annulus, muscle, joints, ligaments). These tissues are innervated primarily by nerve branches of the dorsal ramus. Because typically there is no known pain-provoking pathology in the leg in LBP, leg pain is presumed to reflect spread from the primary source in the back—that is, somatic referred pain [4,16]. The neuropathy hypothesis, in contrast, holds that leg pain is a consequence of ectopic impulse discharge (ectopia) generated paraspinally in compressed or irritated ventral ramus afferents—that is, in sensory axons of the spinal nerves and roots that innervate the leg, and/or in their cell bodies in the corresponding dorsal root ganglia (DRG). According
http://dx.doi.org/10.1016/j.pain.2014.09.015 0304-3959/Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Ò
2552
R. Defrin et al. / PAIN 155 (2014) 2551–2559
to this hypothesis, back pain may have 2 causes: neuropathic ectopia in injured dorsal ramus afferents or sensitized nociceptor endings in deep back tissues. In most patients, both mechanisms probably contribute, yielding qualitatively different pain sensations [2] with different secondary consequences and treatment profiles. Current pain theory holds that sustained peripheral noxious input, whether due to sensitized sensory endings or ectopic pacemaker activity, may secondarily initiate a state of spinal central sensitization [33,41,46]. In this state, afferent input is amplified and activity in low threshold Ab mechanosensitive afferents is rendered painful (Ab pain). A well-known example is secondary hyperalgesia, a region of hypersensibility to light touch (tactile allodynia) on the skin that surrounds the location of a primary noxious input. The central sensitization theory makes a testable prediction. If the primary generators of back pain are nociceptor endings in deep back tissues (territory of the dorsal ramus), patients ought to show tactile allodynia on the skin of the low back but not in the leg. However, if there is neuropathic ectopia in ventral ramus afferents, tactile allodynia should occur in the cutaneous projection field of these afferents, ie, on skin of the leg and foot. We tested this prediction, with results that endorse the neuropathy model in LBP patients with radicular symptoms. These results also validate the common use of hind paw tactile allodynia as a surrogate marker for leg pain in animal research models of LBP. 2. Methods 2.1. Subjects We examined 74 subjects, 35 men and 39 women, with chronic LBP with radicular radiation into the leg (shooting pain). All reported pain in the back and in the leg. In about half of the subjects, the radiating leg pain extended as far as the foot (38 of 74, 51%). In the remainder, it stopped around the ankle (18 of 74, 24%), somewhat below the knee (16 of 74, 22%), or at the knee (2 of 74, 3%). Shooting or radiating pain typically involved a sense of movement of the pain in a proximal-to-distal direction down the course of the sciatic nerve (hence the term sciatica). Back pain was usually bilateral (44 of 74), radicular pain less so (24 of 74, v2 P = .002). In addition, we tested 15 patients with axial LBP (all bilateral, no pain reported in any part of either leg) and 22 healthy controls. Finally, to get a feel for the intensity of tactile allodynia in LBP patients in comparison with a condition in which this symptom is typically a primary cause of suffering, we evaluated a convenience sample of 5 patients who attended the pain clinic who had chronic regional pain syndrome (CRPS type 1) in association with minor trauma. Subject demographics are listed in Table 1. A diagnosis of axial LBP, made independently by the referring physician and a pain specialist, required the following: (1) significant aching and/or stabbing pain felt in the lower back that had been present for P3 months, and (2) a diagnosis of lumbar spine pathology based on clinical examination supported by positive CT and/or MRI scans showing disc protrusion (other spinal
abnormalities were also seen occasionally). A diagnosis of LBP with radicular leg pain additionally required (3) pain with a dermatomal distribution that radiated in a distal direction along the course of the sciatic nerve at least to the knee, and (4) that the patient’s typical radiating pain could be provoked by straight-leg raising (Lasègue sign). About 60% of the patients also had negative signs of sensory root dysfunction such as feelings of numbness and/or motor root dysfunction such as weakness or diminished quadriceps femoris or triceps surae reflexes, supported by abnormal nerve conduction studies or electromyography. Exclusion criteria were the following: (1) evidence of skin disease, (2) conditions with a potential neurological cause or consequences (eg, Parkinsonism, diabetes, central nervous system injury), (3) pain originating deep in the leg potentially due to thrombosis or deep tissue injury, (4) psychiatric illness, and (5) inability to communicate and to understand instructions. The radicular patients all fell into the category of Treede et al. [42] of ‘‘probable neuropathic pain.’’ We note, however, that we tested quantitatively for impaired tactile detection threshold only in a minority of the patients. Had we done this systematically and found striking deficits, classification as ‘‘definite neuropathic pain’’ might have been justified for some of the patients. The patients with CRPS, who all had pain in 1 leg for at least 1 year, met the clinical diagnostic criteria of Harden et al. [15]. Patients were recruited from among attendees of the Unit for Pain Control of the Tel Aviv–Sourasky Medical Center. Most (>90%) were receiving analgesic medications at the time of testing, mostly nonsteroidal anti-inflammatory drugs but some opiates. None had undergone prior exploration for tactile allodynia. Healthy controls were volunteers recruited from among the students and workers of Tel-Aviv University. None were paid for participating. Informed consent was obtained from all participants. The study was approved by the Helsinki committee of the Tel Aviv– Sourasky Medical Center. Subjects were unaware of the objectives of the study. The individuals who carried out the sensory testing were aware that the study involved pain mechanisms and that we were examining the possibility of tactile allodynia being present in some individuals with LBP, but they were not briefed on the theoretical background or the expectations of the principal investigators. 2.2. Study design We searched for allodynia of various types in 1 leg in all 116 subjects (the more painful leg in bilateral patients) and for allodynia on the lower back as well in 61 of them (34 consecutive radicular patients, all 15 patients with axial LBP, and 12 controls). In 22 radicular patients, the spatial location of allodynia was mapped in particular detail and on both sides of the body. These included a run of 15 consecutive patients who were otherwise like the full cohort, and 7 patients selected for having bilateral radicular pain (hence 29 painful legs). Of these 7, 4 had radiological evidence of spinal stenosis in lower lumbar segments (2 at L4–5, 2 at L4–S1). Because these vertebrae are
Table 1 Subject demographics. Group
n
Sex, M/F
Age, y, mean ± SD
Pain duration, y, mean ± SD
LBP with radiation
74
35/39
65.8 ± 12.9
Axial LBP CRPS Healthy controls
15 5 22
9/6 2/3 10/12
64.5 ± 20.7 35.1 ± 7.9 54.2 ± 18.6
4.5 ± 5.5 (leg) 8.8 ± 15.8 (back) 9.2 ± 9.7 (back) 5.2 ± 2.0 (leg)
LBP = low back pain; CRPS = chronic regional pain syndrome.
Ò
R. Defrin et al. / PAIN 155 (2014) 2551–2559
>12 cm caudal to the lumbar enlargement, spinal roots central to the DRG (cauda equina) would have been affected, not the spinal cord. There were no clinical or radiological indications of central pain in any of the patients studied. Testing was done in a quiet room in a single session usually lasting 45 to 60 min. Subjects were blindfolded or asked to close their eyes, and they either sat or lay (according to their preference) in a comfortable supine position. On both the symptomatic and the contralateral leg and back, we tested for the presence of cutaneous tactile sensibility in the dermatome corresponding to the most severe radiologically defined spinal pathology (primary dermatome) and also in the 2 adjacent dermatomes. These were usually L3–L5 but were occasionally L2–L4 or L4–S1. The distribution of primary dermatomes for the radicular patients was: L2 1/74, L3 10/74, L4 29/74, L5 32/74, S1 2/74 (divided proportionately among patients with and without allodynia); for the axial pain patients: L3 2/15, L4 4/15, L4 and 5 5/15, L5 3/15, S1 1/15. Each of 5 types of test stimuli was applied once at a set of standard sites along leg dermatomes and on the lower back (about 5 cm off the midline). Dermatomes overlap broadly [14]. For the present purposes, they were defined using the dermatomal map of Keegan and Garrett [21] as rendered by Netter [30] (Fig. 1). After each stimulus, the subject was asked to state if pain was felt (yes or no, forced choice) with 1 repeat permitted in the event of uncertainty. If pain was reported the subject, still blindfolded,
2553
was asked to point a finger at the pain source. Pain intensity was assessed by a visual analog scale (VAS) with 0 designating no pain and 10 designating the most intense pain imaginable. The eyes were uncovered for VAS rating. A leg dermatome and/or low-back test site was scored as positive for allodynia of a particular type if the corresponding stimulus evoked a pain response. Test stimuli were designed to detect 5 qualities of allodynia. In the calibration stage, they were adjusted to be uniformly nonpainful in 8 healthy volunteers, and this was subsequently reconfirmed in 14 more. The testing order of the dermatomes was randomized as were the 5 test stimuli. The stimuli were as follows. (1) In von Frey–evoked static tactile allodynia, a von Frey monofilament (#4.74 in the Semmes-Weinstein Touch-Test; North Coast Medical Inc.) was applied for 1 s perpendicular to the skin at just bending force (58.9 mN; 6.0 gram-force [48] (1 s von Frey). (2) von Frey– evoked dynamic tactile allodynia was like (1) except the monofilament was minimally bowed and drawn across the skin surface in a 4 cm (1 s) distal-to-proximal stroke (dynamic von Frey) [35]. Perpendicular force, measured on a pan balance, was 1.7 mN (0.17 g). For stimuli (1) and (2), a single monofilament was used for all subjects. (3) In dynamic brush-evoked tactile allodynia, a soft toothbrush (Oral B; Procter and Gamble Co.) was oriented such that all bristles touched the skin (nylon bristles; 1 cm wide, 3 cm long). It was then gently drawn across the skin in a distal-to-proximal direction in a 4 cm stroke lasting 1 s (dynamic brush) [2,35].
Fig. 1. Dermatomal map following Netter [30] showing regions in which allodynia was reported in patients with radicular LBP (red, left leg of figurine) and axial LBP (blue, right leg of figurine). Percentages marked for each region indicate the proportion of legs tested on which allodynia of 1 or more types was reported. Data are based on reports of 15 patients with axial LBP (the more prominent side when bilateral) and 22 patients with radicular LBP (7 bilateral, hence a percentage of 29 legs). Percentages on the front of the leg (including the dorsal and plantar foot) are based on measurements in the L5 dermatome. Percentages on the back of the leg are based on measurements in the S1 dermatome.
Ò
2554
R. Defrin et al. / PAIN 155 (2014) 2551–2559
The force applied normal to the skin was measured as 15.7 mN (1.6 g). (4) Dynamic cotton wool–evoked tactile allodynia was like (3), except that the brush was replaced with a ball of cotton wool about 3 cm in diameter (dynamic cotton) [23,27]. The force applied normal to the skin was measured at 1.6 mN (0.16 g). (5) For coldevoked allodynia, the skin was gently touched for 1 s with a ball of cotton wool moistened with 95% ethanol. Evaporative cooling briefly lowered the skin temperature to 18 °C. In the 22 subjects explored with a particular focus on pain localization, a few changes were made in the testing protocol, as follows. (1) A second von Frey–evoked static tactile allodynia stimulus was added, identical to the first except that the bowed monofilament was held in place for 7 s rather than 1 s (7 s von Frey). The aim was to emulate the corresponding test as it is usually applied in animal models of tactile allodynia [7]. (2) To maintain the original number of stimuli, we dropped the dynamic cotton stimulus because it correlated with the other tests of dynamic tactile allodynia (dynamic von Frey P = .002; dynamic brush P = .001). (3) The dorsal and plantar surfaces of the foot were tested (Fig. 1). We also measured the threshold for detecting light touch in allodynic and contralateral normal skin in 12 of these patients and in a sample of 10 axial LBP patients, the 5 CRPS patients and 8 controls, with the 1 s von Frey stimulus on the midcalf. For this, we used an ascending staircase method that used 1 s applications of a series of Semmes-Weinstein monofilaments (0.08, 0.20, 0.39, 0.69, 1.57, 3.92 mN; corresponding to 0.01, 0.02, 0.04, 0.07, 0.16, 0.40 g). We began with the monofilament with the weakest bending force (0.08 mN). If this stimulus was not detected (no responses to 3 repetitions at 10 s intervals) we proceeded to the next stiffest monofilament, and so forth. The detection threshold was defined as the weakest monofilament detected on at least 2 of the 3 repetitions and correctly located. Before sensory testing, all patients with LBP (with or without sciatica) were interviewed regarding their typical back and leg pain over the previous 30 days. Recorded were the pain’s location (plotted on a body chart), quality, first pain appearance in back or leg, duration from onset, alleviating and exacerbating factors, movement limitations, and medications used. Because LBP tends to fluctuate over time, we obtained maximum and minimum values. We also obtained a global VAS rating of typical pain (not during flareups) over the past month in response to the questions, ‘‘State the intensity of your everyday pain in the back’’ and ‘‘. . .in the leg.’’ Finally, patients filled out the full McGill Pain Questionnaire (MPQ; validated Hebrew translation by M. Weisenberg, Bar Ilan University, Ramat Gan, Israel), either before or after sensory test-
ing. MPQ scores capture patients’ pain experience on sensory, affective, and cognitive dimensions based on words chosen from a standardized list [28]. Four parameters were derived: (1) the number of words chosen, (2) the pain rating index (the sum of the intensity values of the words chosen), and (3) the overall pain magnitude (1–5 scale) when pain is at its least (Pleast) and (4) at its most (Pworst). 2.3. Statistical evaluation Data were analyzed by PASWS Statistics software v17.1 (NCSS). Mean values of continuous variables are provided as ±standard deviation (SD), or alternatively the median (for Pleast and Pworst). Group differences were evaluated by Student’s t tests or the Mann-Whitney test, respectively. Categorical variables described by frequency, and between-group differences were evaluated by the chi-square test or Fisher’s exact probabilities test. The Spearman correlation coefficient was used to detect associations among pain questionnaire parameters and scores from the quantitative sensory testing. Twotailed P values of 6.05 after correction for multiple comparisons using false discovery rate statistics were considered significant. 3. Results 3.1. Allodynia on the leg More than half of the subjects with radicular pain radiating into the leg reported feeling allodynia on the symptomatic leg (45 of 74, 60.8%). There was no sex difference (P = .64). The location of allodynia always matched the location of the stimulus. Skin of the primary dermatome, the dermatome with the greatest pathology as defined radiologically, was only slightly more responsive than the 2 adjacent dermatomes. The likelihood of reporting allodynia varied with the specific test stimulus (Table 2). In contrast to the radicular pain patients, few patients with axial LBP reported any kind of allodynia anywhere on the leg (2 of 15, 13.3%, v2 P < .001), nor was allodynia reported by any of the healthy controls. The proportion of patients receiving analgesic medications did not differ in radicular patients with vs without leg allodynia (v2 P > .05), nor did it differ between (all) radicular patients and the axial LBP patients (v2 P = .5). It is likely that the overall prevalence and/or intensity of allodynia would have been even higher had more of the patients not been receiving analgesics, but we cannot state this with certainty. Nearly all of the 45 radicular patients who experienced allodynia (41 of 45, 91%) reported pain in response to more than 1 of
Table 2 Prevalence of various types of allodynia in the primary dermatome on the leg and lower back.a Stimulus
Static 1 s Static 7 s Dynamic von Frey Dynamic brush Dynamic cotton wool Cold Any type
LBP with radiation, allodynia in leg
Axial LBP, allodynia in leg
LBP with radiation, allodynia in lower back
Axial LBP, allodynia in lower back
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
31/45 (68.8) 14/16 (87.5) 26/45 (57.7)
3.9 ± 2.5 4.5 ± 1.9 4.7 ± 2.1
1/15 (6.7) 1/15 (6.7) 0/15 (0)
4 8
8/34 (23.5) 7/22 (31.8) 5/34 (14.7)
4.6 ± 3.3 5.6 ± 2.3 4.8 ± 3.2
7/15 (46.6) 2/5 (40.0) 2/15 (13.3)
6.2 ± 1.8 6.5 ± 2.1 6.5 ± 0.7
22/45 (48.8) 16/29 (55.1)
4.4 ± 2.5 4.0 ± 1.9
0/15 (0) 0/15 (0)
1/34 (2.9) 1/12 (8.3)
1.0 2.0
2/15 (13.3) 1/9 (11.1)
5.0 ± 2.8 2.0
26/45 (57.7) 45/74 (60.8)
3.7 ± 2.3 4.2 ± 2.3
0/15 (0) 2/15 (13.3)
6/34 (17.6) 10/34 (29.4)
3.6 ± 2.6 3.6 ± 1.8
5/15 (33.3) 8/15 (53.3)
2.2 ± 0.8 4.8 ± 2.2
6.0 ± 2.8
LBP = low back pain; VAS = visual analog scale. a Intensity is rated on a 0–10 point VAS scale in patients with radicular low back pain (with radiation) and in patients with LBP without radiation (axial LBP). The denominator for calculating percentages is not identical for all stimuli.
Ò
2555
LBP = low back pain; VAS = visual analog scale. a Pain was mapped in finer detail in these patients than in the others and on both sides of the body. Allodynia was reported by 16 of these individuals, bilaterally in 7. Thus, data from a total of 23 legs are represented. Values are the number of testing sites (n/23) at which the various types of allodynia were reported, and its average intensity rated on a 0–10 point VAS.
4.0 ± 2.6 4.3 ± 2.2 12/23 (39.1) 23/23 (100) 4.7 ± 2.7 5.1 ± 1.4 7 (30.4) 7/23 (30.4) 4.3 ± 2.7 4.5 ± 2.0 3.5 ± 2.1 4.1 ± 1.7 4.1 ± 2.7 4.6 ± 2.7 7 (30.4) 10/23 (43.5)
7 (30.4) 12/23 (52.2)
3.8 ± 2.5 4.7 ± 2.3
6 (26.1) 19/23 (82.6)
3.6 ± 3.1 4.2 ± 2.1
7 (30.4) 9/23 (11.9)
6 (26.1) 8/23 (34.8)
4.4 ± 2.3 9/23 (11.9) 5.5 ± 0.6 4 (17.4) 2 (8.7) 4.0 ± 4.2 2 (8.7)
2.5 ± 2.1
3 (13.1)
3.6 ± 2.5
4 (17.4) 3.7 ± 1.9 4 (17.4)
5 ± 1.6
4.1 ± 2.3 4.3 ± 2.1 4.9 ± 1.7 12/23 (52.2) 20/23 (86.9) 13/23 (56.5) 5.0 5 ± 1.0 4.7 ± 1.1 1 (4.3) 3 (13.1) 3 (13.1)
Static 1 s Static 7 s Dynamic von Frey Dynamic brush Cold Any type
4 (17.4) 7 (30.4) 8 (43.8) 5.5 ± 2.8 3.7 ± 2.5 5.2 ± 2.7 6 (26.1) 7 (30.4) 6 (26.1)
3.4 ± 1.8 5.7 ± 2.3 5.8 ± 2.3
5 (21.7) 14 (60.9) 8 (43.8)
3.2 ± 1.9 4.6 ± 1.8 5.1 ± 1.8
3 (13.1) 8 (43.8) 6 (26.1) 1.0 4.6 ± 2.1 4.7 ± 0.9 1 (4.3) 6 (26.1) 4 (17.4)
5.0 ± 1.0 4.0 ± 1.7 4.3 ± 1.9
Any location Plantar foot Dorsal foot
n/23 (%) VAS ± SD
Midcalf Lower thigh
n/23 (%) n/23 (%)
VAS ± SD Upper thigh
n/23 (%) VAS ± SD
Lower back
Type of allodynia
Table 3 Types and locations of allodynia in the primary dermatome in 22 radicular pain patients.a
VAS ± SD
n/23 (%)
VAS ± SD
n/23 (%)
VAS ± SD
n/23 (%)
VAS ± SD
R. Defrin et al. / PAIN 155 (2014) 2551–2559
the 5 types of nonnoxious stimuli evaluated, averaging 2.8 ± 1.6 types of allodynia per patient (range 1–5). Considering only the primary dermatome, pain was reported to static tactile stimuli by 73.8% of the patients (1 s von Frey 31 of 45; 7 s von Frey 14 of 16, v2 P > .1). This was followed by cold-evoked allodynia in 57.7% (26 of 45) and dynamic tactile allodynia in 53.8% (dynamic von Frey, 26 of 45; dynamic cotton, 16 of 29; dynamic brush 22 of 45; Table 2). Cold was painful as frequently as tactile stimulation (P > .1). Pain in response to the 7 s von Frey static tactile stimulus was more prevalent than to each of the 3 dynamic tactile stimuli (P < .05). This was surprising in light of the typical predominance of dynamic tactile allodynia in neuropathy. This result was not obtained for the 1 s von Frey static stimulus (P = .382, 0.343 and 0.087, respectively). The 2 axial LBP patients who reported leg allodynia each responded to only 1 stimulus, the 1 s or 7 s static tactile stimulus (Table 2). As in the radicular pain patients at large, most of those who were mapped closely reported allodynia (16 of 22, 72.7%). This included all 7 of the bilateral patients tested (thus 23 of 29 legs, 79.3%), allowing us to compare the 2 legs in the same individual. The 5 types of stimuli used were delivered at 6 locations—low back, upper thigh, lower thigh, midcalf, dorsal foot, and plantar foot— with 30 tests per dermatome (Fig. 1). Nearly one quarter of these nonnoxious stimuli were reported as painful in the 16 allodynic patients (6.9 ± 6.3 per dermatome in 23 legs). Tactile allodynia was reported on the foot (dorsal and/or plantar) in nearly half of the subjects (7 of 16 patients, 9 of 23 legs). Numbers were similar for cold allodynia (6 of 16 patients, 9 of 23 legs). For any type of allodynia, tactile or cold, the numbers were 8 of 16 patients and 12 of 23 legs. The ranking of allodynia types was similar to the larger group of patients (Tables 2 and 3). In the 7 bilateral patients, allodynia was similar in quality on the 2 legs, with a similar number of allodynia types (P = .18). 3.2. Allodynia on the lower back Radicular patients also reported feeling allodynia on skin of the lower back, although less frequently than on the leg [29.4% of patients tested on the back (10 of 34) vs the 60.8% (45 of 74) who reported leg allodynia; v2 P = .005; Table 4, Fig. 1]. In patients with axial LBP, allodynia on the back occurred at about the same frequency (8 of 15 vs 10 of 34; P = .12). Thus, having pain in the leg did not alter the likelihood of having allodynia on the back. The much larger size of the ventral ramus may explain the predominance of leg allodynia over back allodynia in patients with ventral ramus involvement (ie, radicular patients). Back pain was bilateral in all of the axial LBP patients, and in those who had allodynia on the back, the allodynia was also usually bilateral (7 of 8, 88%). Bilateral back pain was less common in the radicular patients (20 of 34, 56.8%; P = .002), but when it occurred, allodynia on the back was also usually bilateral (8 of 10, 80%). Thus, while allodynia did not always accompany leg or back pain, it only occurred when adjacent underlying deep background pain was present on the same side. There were no instances of mirror allodynia. Allodynia on the leg and back were similar in
Table 4 Proportion of patients tested who had LBP with radiation or axial LBP (LBP without radiation) and reported allodynia on the leg, lower back, or both in the primary dermatome.
a
Location of allodynia
LBP with radiation
Axial LBP
P valuea
Leg Lower back
45/74 (60.8%) 10/34 (29.4%)
2/15 (13.3%) 8/15 (53.3%)
.001 .12
Statistical tests are based on v2 or Fisher’s exact probabilities tests.
Ò
2556
R. Defrin et al. / PAIN 155 (2014) 2551–2559
Table 5 Force threshold for first detecting a light touch stimulus (1 s von Frey) in allodynic skin on the back and midcalf.a Group
n
Back, mN ± SD
Painful leg, mN ± SD
Nonpainful leg, mN ± SD
LBP with radiation Axial LBP
12 10
0.39 ± 0.3 1.3 ± 2.2
1.1 ± 1.6
5 8
Not tested 0.2 ± 0.06
1.05 ± 1.4
1.0 ± 1.7 1.9 ± 2.1 1.4 ± 1.5 1.1 ± 1.3 0.5 ± 0.1 0.5 ± 0.2
CRPS (unilateral) Healthy controls
LBP = low back pain; CRPS = chronic regional pain syndrome. a Corresponding values are provided for the contralateral calf at the mirror-image location on the nonpainful leg. When neither leg was painful (axial LBP and for healthy controls), values for both legs are given under ‘‘nonpainful leg.’’ None of the values differed significantly (P > .05).
quality and intensity (Tables 2 and 3). None of the control subjects reported allodynia on the back. 3.3. Intensity Like other cohorts of radicular pain patients surveyed [26], our patients seldom complained of the presence of tactile or cold allodynia during the interviews or in the MPQ, although other spontaneous sensations such as pins and needles, electrical currents and numbness were mentioned. So as not to draw attention to such symptoms, we did not ask to what extent they contributed to the overall burden of pain. When present, allodynia on the leg and back was rated as moderate (about 4 on the 0–10 VAS; Tables 2 and 3). But even at its worst, LBP patients rated their tactile allodynia as much milder than CRPS patients (P < .001), who uniformly gave VAS scores of 10. Interestingly, the tactile detection threshold in LBP patients, CRPS patients, and healthy controls was similar (Table 5) even though the quality of the sensation reported at threshold stimulus intensity was very different (pain vs touch). This suggests that in all cases, the sensation at threshold was signaled by the same type of low-threshold mechanoreceptive afferents [33,41,46]. Global VAS intensity ratings and MPQ scales measured the patients’ overall pain experience over the past month. The radicular patients uniformly reported more intense and persistent overall pain than those with axial LBP (Table 6), and the intensity and duration of their radicular (leg) pain correlated with that of their back pain (P < .001, Table 7). Likewise, radicular patients with leg allodynia reported more deep back and leg pain than those without on most scales (P < .05, Table 6). A striking exception was the global VAS measure of leg pain intensity, where there was no difference (P = .25). Apparently the development of tactile allodynia, and by inference of central sensitization, has more to do with individual predisposition than with the intensity of the precipitating noxious input. However, when allodynia was present, the global
intensity of leg pain did correlate significantly with the intensity of the allodynia, for 2 subtypes at least: dynamic von Frey and dynamic cotton (P = .01 and .05 respectively, Table 7). Although longitudinal data were not collected, one can infer from the time-series data of patients with progressively longerlasting symptoms that allodynia likely increases over time. Specifically, the prevalence of static tactile allodynia correlated positively with the duration of radicular pain (P = .04) and that of cold allodynia trended in the same direction (P = .06, Table 7). Remarkably, however, there appeared to be a negative relation between intensity (VAS) rating of cold allodynia and the duration of radicular pain (P = .035) and a hint of such a relation also with static tactile allodynia (P = .066, Table 7). This was also true for back pain (P < .01 and .02). Our tentative inference is that as time passed, tactile and cold allodynia faded in intensity just as they were becoming more common. Prospective studies are required to confirm that these static observations indeed represent dynamic processes that evolve over time. The results in hand, however, suggest that allodynia emerges with a delay after the appearance of radicular pain, but with its intensity near maximum at the time of emergence. Subsequently it tends to fade. If this pattern is confirmed by longitudinal data in individual patients followed over time, bedside tracking of the first appearance of tactile and cold allodynia, as well as of progressive changes over time, might provide valuable prognostic information on the likely course of the condition [48]. 4. Discussion The classic signs and symptoms of LBP are well characterized [4,12]. Here we documented the frequent presence of a novel sign, cutaneous tactile allodynia. The sole comparable study [13] tested a single stimulus modality (dynamic mechanical) at a single leg location (foot dorsum). One of 12 patients tested reported marked allodynia. The authors, however, dismissed this case as an outlier. In fact, their result was no different from ours at this test location
Table 6 Comparison of various pain measures in LBP patients with and without pain radiation into the leg.a Pain measure
MPQ-PRI MPQ-NWC MPQ-Pleast MPQ-Pworst Radicular pain (VAS) Back pain (VAS) Radicular pain (duration, y) Back pain duration, y)
LBP with radiation
Axial LBP
Comparison, P value
Allodynia (n = 45)
No allodynia (n = 29)
Allodynia (n = 2)
No allodynia (n = 13)
30.8 ± 14.8⁄ 11.8 ± 5.6⁄ 2 5 6.3 ± 2.9 6.7 ± 1.9⁄⁄ 5.7 ± 5.6⁄⁄ 7.9 ± 7.5
23.2 ± 12.7 9.5 ± 4.8 2 5 5.6 ± 2.7 4.9 ± 3.0 2.7 ± 2.9 10.6 ± 14.7
6.5 ± 0.8 9.1 ± 5.9 2 4 0 4.2 ± 0.3
16.5 ± 13.6 8.1 ± 7.8 1.5 4 0 6.4 ± 1.9
PAIN 155 (2014) 2551–2559
www.elsevier.com/locate/pain
Tactile allodynia in patients with lumbar radicular pain (sciatica) Ruth Defrin a, Marshall Devor b,⇑,1, Silviu Brill c,1 a
Department of Physical Therapy, School of Allied Health Professions, The Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel Department of Cell and Developmental Biology, Institute of Life Sciences, and Center for Research on Pain, The Hebrew University of Jerusalem, Jerusalem 91904, Israel c Institute of Pain Medicine, Department of Anesthesia and Critical Care Medicine, Tel Aviv Sourasky Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel-Aviv 69978, Israel b
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
a r t i c l e
i n f o
Article history: Received 25 March 2014 Received in revised form 28 August 2014 Accepted 11 September 2014
Keywords: Central sensitization Radicular pain Sciatica Shooting pain Tactile allodynia
a b s t r a c t We report a novel symptom in many patients with low back pain (LBP) that sheds new light on the underlying pain mechanism. By means of quantitative sensory testing, we compared patients with radicular LBP (sciatica), axial LBP (LBP without radiation into the leg), and healthy controls, searching for cutaneous allodynia in response to weak tactile and cooling stimuli on the leg and low back. Most patients with radicular pain (~60%) reported static and dynamic tactile allodynia, as well as cooling allodynia, on the leg, often extending into the foot. Some also reported allodynia on the low back. In axial LBP, allodynia was almost exclusively on the back. The degree of dynamic tactile allodynia correlated with the degree of background pain. The presence of allodynia suggests that the peripheral nerve generators of background leg and back pain have also induced central sensitization. The distal (foot) location of the allodynia in patients who have it indicates that the nociceptive drive that maintains the central sensitization arises paraspinally (ectopically) in injured ventral ramus afferents; this is not an instance of somatic referred pain. The presence of central sensitization also provides the first cogent account of shooting pain in sciatica as a wave of activity sweeping vectorially across the width of the sensitized dorsal horn. Finally, the results endorse leg allodynia as a pain biomarker in animal research on LBP, which is commonly used but has not been previously validated. In addition to informing the underlying mechanism of LBP, bedside mapping of allodynia might have practical implications for prognosis and treatment. Social media question: How can you tell whether pain radiating into the leg in a patient with sciatica is neuropathic, ie, due to nerve injury? Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
1. Introduction Low back pain (LBP) is consistently ranked among the leading causes of disability due to disease or injury [44]. Treatment is often unsatisfactory, at least in part as a result of uncertainty concerning the underlying mechanisms. In particular, radiological evidence of spinal pathology correlates imperfectly with pain symptoms [4,12,35] suggesting that it is not the pathology per se that triggers pain, but rather secondary changes in neural function that are triggered by the pathology in susceptible individuals. Here we report new observations on symptomatology that shed light on LBP mechanisms and that may have direct implications for diagnosis, prognosis, and treatment. ⇑ Corresponding author. E-mail address: [email protected] (M. Devor). The last 2 authors contributed equally to this article, and both should be considered senior author. 1
LBP, which frequently also features pain in the leg (lumbar radicular pain, or sciatica), is variously considered to reflect the activation of nociceptive sensory endings (normal or sensitized by inflammation), or a neuropathic process in the peripheral nervous system [1,4,8,10,12,13,22,31]. According to the nociception/ inflammation hypothesis, pain results from electrical impulses originating in sensory endings in deep peripheral tissues (eg, disc, annulus, muscle, joints, ligaments). These tissues are innervated primarily by nerve branches of the dorsal ramus. Because typically there is no known pain-provoking pathology in the leg in LBP, leg pain is presumed to reflect spread from the primary source in the back—that is, somatic referred pain [4,16]. The neuropathy hypothesis, in contrast, holds that leg pain is a consequence of ectopic impulse discharge (ectopia) generated paraspinally in compressed or irritated ventral ramus afferents—that is, in sensory axons of the spinal nerves and roots that innervate the leg, and/or in their cell bodies in the corresponding dorsal root ganglia (DRG). According
http://dx.doi.org/10.1016/j.pain.2014.09.015 0304-3959/Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Ò
2552
R. Defrin et al. / PAIN 155 (2014) 2551–2559
to this hypothesis, back pain may have 2 causes: neuropathic ectopia in injured dorsal ramus afferents or sensitized nociceptor endings in deep back tissues. In most patients, both mechanisms probably contribute, yielding qualitatively different pain sensations [2] with different secondary consequences and treatment profiles. Current pain theory holds that sustained peripheral noxious input, whether due to sensitized sensory endings or ectopic pacemaker activity, may secondarily initiate a state of spinal central sensitization [33,41,46]. In this state, afferent input is amplified and activity in low threshold Ab mechanosensitive afferents is rendered painful (Ab pain). A well-known example is secondary hyperalgesia, a region of hypersensibility to light touch (tactile allodynia) on the skin that surrounds the location of a primary noxious input. The central sensitization theory makes a testable prediction. If the primary generators of back pain are nociceptor endings in deep back tissues (territory of the dorsal ramus), patients ought to show tactile allodynia on the skin of the low back but not in the leg. However, if there is neuropathic ectopia in ventral ramus afferents, tactile allodynia should occur in the cutaneous projection field of these afferents, ie, on skin of the leg and foot. We tested this prediction, with results that endorse the neuropathy model in LBP patients with radicular symptoms. These results also validate the common use of hind paw tactile allodynia as a surrogate marker for leg pain in animal research models of LBP. 2. Methods 2.1. Subjects We examined 74 subjects, 35 men and 39 women, with chronic LBP with radicular radiation into the leg (shooting pain). All reported pain in the back and in the leg. In about half of the subjects, the radiating leg pain extended as far as the foot (38 of 74, 51%). In the remainder, it stopped around the ankle (18 of 74, 24%), somewhat below the knee (16 of 74, 22%), or at the knee (2 of 74, 3%). Shooting or radiating pain typically involved a sense of movement of the pain in a proximal-to-distal direction down the course of the sciatic nerve (hence the term sciatica). Back pain was usually bilateral (44 of 74), radicular pain less so (24 of 74, v2 P = .002). In addition, we tested 15 patients with axial LBP (all bilateral, no pain reported in any part of either leg) and 22 healthy controls. Finally, to get a feel for the intensity of tactile allodynia in LBP patients in comparison with a condition in which this symptom is typically a primary cause of suffering, we evaluated a convenience sample of 5 patients who attended the pain clinic who had chronic regional pain syndrome (CRPS type 1) in association with minor trauma. Subject demographics are listed in Table 1. A diagnosis of axial LBP, made independently by the referring physician and a pain specialist, required the following: (1) significant aching and/or stabbing pain felt in the lower back that had been present for P3 months, and (2) a diagnosis of lumbar spine pathology based on clinical examination supported by positive CT and/or MRI scans showing disc protrusion (other spinal
abnormalities were also seen occasionally). A diagnosis of LBP with radicular leg pain additionally required (3) pain with a dermatomal distribution that radiated in a distal direction along the course of the sciatic nerve at least to the knee, and (4) that the patient’s typical radiating pain could be provoked by straight-leg raising (Lasègue sign). About 60% of the patients also had negative signs of sensory root dysfunction such as feelings of numbness and/or motor root dysfunction such as weakness or diminished quadriceps femoris or triceps surae reflexes, supported by abnormal nerve conduction studies or electromyography. Exclusion criteria were the following: (1) evidence of skin disease, (2) conditions with a potential neurological cause or consequences (eg, Parkinsonism, diabetes, central nervous system injury), (3) pain originating deep in the leg potentially due to thrombosis or deep tissue injury, (4) psychiatric illness, and (5) inability to communicate and to understand instructions. The radicular patients all fell into the category of Treede et al. [42] of ‘‘probable neuropathic pain.’’ We note, however, that we tested quantitatively for impaired tactile detection threshold only in a minority of the patients. Had we done this systematically and found striking deficits, classification as ‘‘definite neuropathic pain’’ might have been justified for some of the patients. The patients with CRPS, who all had pain in 1 leg for at least 1 year, met the clinical diagnostic criteria of Harden et al. [15]. Patients were recruited from among attendees of the Unit for Pain Control of the Tel Aviv–Sourasky Medical Center. Most (>90%) were receiving analgesic medications at the time of testing, mostly nonsteroidal anti-inflammatory drugs but some opiates. None had undergone prior exploration for tactile allodynia. Healthy controls were volunteers recruited from among the students and workers of Tel-Aviv University. None were paid for participating. Informed consent was obtained from all participants. The study was approved by the Helsinki committee of the Tel Aviv– Sourasky Medical Center. Subjects were unaware of the objectives of the study. The individuals who carried out the sensory testing were aware that the study involved pain mechanisms and that we were examining the possibility of tactile allodynia being present in some individuals with LBP, but they were not briefed on the theoretical background or the expectations of the principal investigators. 2.2. Study design We searched for allodynia of various types in 1 leg in all 116 subjects (the more painful leg in bilateral patients) and for allodynia on the lower back as well in 61 of them (34 consecutive radicular patients, all 15 patients with axial LBP, and 12 controls). In 22 radicular patients, the spatial location of allodynia was mapped in particular detail and on both sides of the body. These included a run of 15 consecutive patients who were otherwise like the full cohort, and 7 patients selected for having bilateral radicular pain (hence 29 painful legs). Of these 7, 4 had radiological evidence of spinal stenosis in lower lumbar segments (2 at L4–5, 2 at L4–S1). Because these vertebrae are
Table 1 Subject demographics. Group
n
Sex, M/F
Age, y, mean ± SD
Pain duration, y, mean ± SD
LBP with radiation
74
35/39
65.8 ± 12.9
Axial LBP CRPS Healthy controls
15 5 22
9/6 2/3 10/12
64.5 ± 20.7 35.1 ± 7.9 54.2 ± 18.6
4.5 ± 5.5 (leg) 8.8 ± 15.8 (back) 9.2 ± 9.7 (back) 5.2 ± 2.0 (leg)
LBP = low back pain; CRPS = chronic regional pain syndrome.
Ò
R. Defrin et al. / PAIN 155 (2014) 2551–2559
>12 cm caudal to the lumbar enlargement, spinal roots central to the DRG (cauda equina) would have been affected, not the spinal cord. There were no clinical or radiological indications of central pain in any of the patients studied. Testing was done in a quiet room in a single session usually lasting 45 to 60 min. Subjects were blindfolded or asked to close their eyes, and they either sat or lay (according to their preference) in a comfortable supine position. On both the symptomatic and the contralateral leg and back, we tested for the presence of cutaneous tactile sensibility in the dermatome corresponding to the most severe radiologically defined spinal pathology (primary dermatome) and also in the 2 adjacent dermatomes. These were usually L3–L5 but were occasionally L2–L4 or L4–S1. The distribution of primary dermatomes for the radicular patients was: L2 1/74, L3 10/74, L4 29/74, L5 32/74, S1 2/74 (divided proportionately among patients with and without allodynia); for the axial pain patients: L3 2/15, L4 4/15, L4 and 5 5/15, L5 3/15, S1 1/15. Each of 5 types of test stimuli was applied once at a set of standard sites along leg dermatomes and on the lower back (about 5 cm off the midline). Dermatomes overlap broadly [14]. For the present purposes, they were defined using the dermatomal map of Keegan and Garrett [21] as rendered by Netter [30] (Fig. 1). After each stimulus, the subject was asked to state if pain was felt (yes or no, forced choice) with 1 repeat permitted in the event of uncertainty. If pain was reported the subject, still blindfolded,
2553
was asked to point a finger at the pain source. Pain intensity was assessed by a visual analog scale (VAS) with 0 designating no pain and 10 designating the most intense pain imaginable. The eyes were uncovered for VAS rating. A leg dermatome and/or low-back test site was scored as positive for allodynia of a particular type if the corresponding stimulus evoked a pain response. Test stimuli were designed to detect 5 qualities of allodynia. In the calibration stage, they were adjusted to be uniformly nonpainful in 8 healthy volunteers, and this was subsequently reconfirmed in 14 more. The testing order of the dermatomes was randomized as were the 5 test stimuli. The stimuli were as follows. (1) In von Frey–evoked static tactile allodynia, a von Frey monofilament (#4.74 in the Semmes-Weinstein Touch-Test; North Coast Medical Inc.) was applied for 1 s perpendicular to the skin at just bending force (58.9 mN; 6.0 gram-force [48] (1 s von Frey). (2) von Frey– evoked dynamic tactile allodynia was like (1) except the monofilament was minimally bowed and drawn across the skin surface in a 4 cm (1 s) distal-to-proximal stroke (dynamic von Frey) [35]. Perpendicular force, measured on a pan balance, was 1.7 mN (0.17 g). For stimuli (1) and (2), a single monofilament was used for all subjects. (3) In dynamic brush-evoked tactile allodynia, a soft toothbrush (Oral B; Procter and Gamble Co.) was oriented such that all bristles touched the skin (nylon bristles; 1 cm wide, 3 cm long). It was then gently drawn across the skin in a distal-to-proximal direction in a 4 cm stroke lasting 1 s (dynamic brush) [2,35].
Fig. 1. Dermatomal map following Netter [30] showing regions in which allodynia was reported in patients with radicular LBP (red, left leg of figurine) and axial LBP (blue, right leg of figurine). Percentages marked for each region indicate the proportion of legs tested on which allodynia of 1 or more types was reported. Data are based on reports of 15 patients with axial LBP (the more prominent side when bilateral) and 22 patients with radicular LBP (7 bilateral, hence a percentage of 29 legs). Percentages on the front of the leg (including the dorsal and plantar foot) are based on measurements in the L5 dermatome. Percentages on the back of the leg are based on measurements in the S1 dermatome.
Ò
2554
R. Defrin et al. / PAIN 155 (2014) 2551–2559
The force applied normal to the skin was measured as 15.7 mN (1.6 g). (4) Dynamic cotton wool–evoked tactile allodynia was like (3), except that the brush was replaced with a ball of cotton wool about 3 cm in diameter (dynamic cotton) [23,27]. The force applied normal to the skin was measured at 1.6 mN (0.16 g). (5) For coldevoked allodynia, the skin was gently touched for 1 s with a ball of cotton wool moistened with 95% ethanol. Evaporative cooling briefly lowered the skin temperature to 18 °C. In the 22 subjects explored with a particular focus on pain localization, a few changes were made in the testing protocol, as follows. (1) A second von Frey–evoked static tactile allodynia stimulus was added, identical to the first except that the bowed monofilament was held in place for 7 s rather than 1 s (7 s von Frey). The aim was to emulate the corresponding test as it is usually applied in animal models of tactile allodynia [7]. (2) To maintain the original number of stimuli, we dropped the dynamic cotton stimulus because it correlated with the other tests of dynamic tactile allodynia (dynamic von Frey P = .002; dynamic brush P = .001). (3) The dorsal and plantar surfaces of the foot were tested (Fig. 1). We also measured the threshold for detecting light touch in allodynic and contralateral normal skin in 12 of these patients and in a sample of 10 axial LBP patients, the 5 CRPS patients and 8 controls, with the 1 s von Frey stimulus on the midcalf. For this, we used an ascending staircase method that used 1 s applications of a series of Semmes-Weinstein monofilaments (0.08, 0.20, 0.39, 0.69, 1.57, 3.92 mN; corresponding to 0.01, 0.02, 0.04, 0.07, 0.16, 0.40 g). We began with the monofilament with the weakest bending force (0.08 mN). If this stimulus was not detected (no responses to 3 repetitions at 10 s intervals) we proceeded to the next stiffest monofilament, and so forth. The detection threshold was defined as the weakest monofilament detected on at least 2 of the 3 repetitions and correctly located. Before sensory testing, all patients with LBP (with or without sciatica) were interviewed regarding their typical back and leg pain over the previous 30 days. Recorded were the pain’s location (plotted on a body chart), quality, first pain appearance in back or leg, duration from onset, alleviating and exacerbating factors, movement limitations, and medications used. Because LBP tends to fluctuate over time, we obtained maximum and minimum values. We also obtained a global VAS rating of typical pain (not during flareups) over the past month in response to the questions, ‘‘State the intensity of your everyday pain in the back’’ and ‘‘. . .in the leg.’’ Finally, patients filled out the full McGill Pain Questionnaire (MPQ; validated Hebrew translation by M. Weisenberg, Bar Ilan University, Ramat Gan, Israel), either before or after sensory test-
ing. MPQ scores capture patients’ pain experience on sensory, affective, and cognitive dimensions based on words chosen from a standardized list [28]. Four parameters were derived: (1) the number of words chosen, (2) the pain rating index (the sum of the intensity values of the words chosen), and (3) the overall pain magnitude (1–5 scale) when pain is at its least (Pleast) and (4) at its most (Pworst). 2.3. Statistical evaluation Data were analyzed by PASWS Statistics software v17.1 (NCSS). Mean values of continuous variables are provided as ±standard deviation (SD), or alternatively the median (for Pleast and Pworst). Group differences were evaluated by Student’s t tests or the Mann-Whitney test, respectively. Categorical variables described by frequency, and between-group differences were evaluated by the chi-square test or Fisher’s exact probabilities test. The Spearman correlation coefficient was used to detect associations among pain questionnaire parameters and scores from the quantitative sensory testing. Twotailed P values of 6.05 after correction for multiple comparisons using false discovery rate statistics were considered significant. 3. Results 3.1. Allodynia on the leg More than half of the subjects with radicular pain radiating into the leg reported feeling allodynia on the symptomatic leg (45 of 74, 60.8%). There was no sex difference (P = .64). The location of allodynia always matched the location of the stimulus. Skin of the primary dermatome, the dermatome with the greatest pathology as defined radiologically, was only slightly more responsive than the 2 adjacent dermatomes. The likelihood of reporting allodynia varied with the specific test stimulus (Table 2). In contrast to the radicular pain patients, few patients with axial LBP reported any kind of allodynia anywhere on the leg (2 of 15, 13.3%, v2 P < .001), nor was allodynia reported by any of the healthy controls. The proportion of patients receiving analgesic medications did not differ in radicular patients with vs without leg allodynia (v2 P > .05), nor did it differ between (all) radicular patients and the axial LBP patients (v2 P = .5). It is likely that the overall prevalence and/or intensity of allodynia would have been even higher had more of the patients not been receiving analgesics, but we cannot state this with certainty. Nearly all of the 45 radicular patients who experienced allodynia (41 of 45, 91%) reported pain in response to more than 1 of
Table 2 Prevalence of various types of allodynia in the primary dermatome on the leg and lower back.a Stimulus
Static 1 s Static 7 s Dynamic von Frey Dynamic brush Dynamic cotton wool Cold Any type
LBP with radiation, allodynia in leg
Axial LBP, allodynia in leg
LBP with radiation, allodynia in lower back
Axial LBP, allodynia in lower back
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
n with allodynia/N tested (%)
VAS, mean ± SD
31/45 (68.8) 14/16 (87.5) 26/45 (57.7)
3.9 ± 2.5 4.5 ± 1.9 4.7 ± 2.1
1/15 (6.7) 1/15 (6.7) 0/15 (0)
4 8
8/34 (23.5) 7/22 (31.8) 5/34 (14.7)
4.6 ± 3.3 5.6 ± 2.3 4.8 ± 3.2
7/15 (46.6) 2/5 (40.0) 2/15 (13.3)
6.2 ± 1.8 6.5 ± 2.1 6.5 ± 0.7
22/45 (48.8) 16/29 (55.1)
4.4 ± 2.5 4.0 ± 1.9
0/15 (0) 0/15 (0)
1/34 (2.9) 1/12 (8.3)
1.0 2.0
2/15 (13.3) 1/9 (11.1)
5.0 ± 2.8 2.0
26/45 (57.7) 45/74 (60.8)
3.7 ± 2.3 4.2 ± 2.3
0/15 (0) 2/15 (13.3)
6/34 (17.6) 10/34 (29.4)
3.6 ± 2.6 3.6 ± 1.8
5/15 (33.3) 8/15 (53.3)
2.2 ± 0.8 4.8 ± 2.2
6.0 ± 2.8
LBP = low back pain; VAS = visual analog scale. a Intensity is rated on a 0–10 point VAS scale in patients with radicular low back pain (with radiation) and in patients with LBP without radiation (axial LBP). The denominator for calculating percentages is not identical for all stimuli.
Ò
2555
LBP = low back pain; VAS = visual analog scale. a Pain was mapped in finer detail in these patients than in the others and on both sides of the body. Allodynia was reported by 16 of these individuals, bilaterally in 7. Thus, data from a total of 23 legs are represented. Values are the number of testing sites (n/23) at which the various types of allodynia were reported, and its average intensity rated on a 0–10 point VAS.
4.0 ± 2.6 4.3 ± 2.2 12/23 (39.1) 23/23 (100) 4.7 ± 2.7 5.1 ± 1.4 7 (30.4) 7/23 (30.4) 4.3 ± 2.7 4.5 ± 2.0 3.5 ± 2.1 4.1 ± 1.7 4.1 ± 2.7 4.6 ± 2.7 7 (30.4) 10/23 (43.5)
7 (30.4) 12/23 (52.2)
3.8 ± 2.5 4.7 ± 2.3
6 (26.1) 19/23 (82.6)
3.6 ± 3.1 4.2 ± 2.1
7 (30.4) 9/23 (11.9)
6 (26.1) 8/23 (34.8)
4.4 ± 2.3 9/23 (11.9) 5.5 ± 0.6 4 (17.4) 2 (8.7) 4.0 ± 4.2 2 (8.7)
2.5 ± 2.1
3 (13.1)
3.6 ± 2.5
4 (17.4) 3.7 ± 1.9 4 (17.4)
5 ± 1.6
4.1 ± 2.3 4.3 ± 2.1 4.9 ± 1.7 12/23 (52.2) 20/23 (86.9) 13/23 (56.5) 5.0 5 ± 1.0 4.7 ± 1.1 1 (4.3) 3 (13.1) 3 (13.1)
Static 1 s Static 7 s Dynamic von Frey Dynamic brush Cold Any type
4 (17.4) 7 (30.4) 8 (43.8) 5.5 ± 2.8 3.7 ± 2.5 5.2 ± 2.7 6 (26.1) 7 (30.4) 6 (26.1)
3.4 ± 1.8 5.7 ± 2.3 5.8 ± 2.3
5 (21.7) 14 (60.9) 8 (43.8)
3.2 ± 1.9 4.6 ± 1.8 5.1 ± 1.8
3 (13.1) 8 (43.8) 6 (26.1) 1.0 4.6 ± 2.1 4.7 ± 0.9 1 (4.3) 6 (26.1) 4 (17.4)
5.0 ± 1.0 4.0 ± 1.7 4.3 ± 1.9
Any location Plantar foot Dorsal foot
n/23 (%) VAS ± SD
Midcalf Lower thigh
n/23 (%) n/23 (%)
VAS ± SD Upper thigh
n/23 (%) VAS ± SD
Lower back
Type of allodynia
Table 3 Types and locations of allodynia in the primary dermatome in 22 radicular pain patients.a
VAS ± SD
n/23 (%)
VAS ± SD
n/23 (%)
VAS ± SD
n/23 (%)
VAS ± SD
R. Defrin et al. / PAIN 155 (2014) 2551–2559
the 5 types of nonnoxious stimuli evaluated, averaging 2.8 ± 1.6 types of allodynia per patient (range 1–5). Considering only the primary dermatome, pain was reported to static tactile stimuli by 73.8% of the patients (1 s von Frey 31 of 45; 7 s von Frey 14 of 16, v2 P > .1). This was followed by cold-evoked allodynia in 57.7% (26 of 45) and dynamic tactile allodynia in 53.8% (dynamic von Frey, 26 of 45; dynamic cotton, 16 of 29; dynamic brush 22 of 45; Table 2). Cold was painful as frequently as tactile stimulation (P > .1). Pain in response to the 7 s von Frey static tactile stimulus was more prevalent than to each of the 3 dynamic tactile stimuli (P < .05). This was surprising in light of the typical predominance of dynamic tactile allodynia in neuropathy. This result was not obtained for the 1 s von Frey static stimulus (P = .382, 0.343 and 0.087, respectively). The 2 axial LBP patients who reported leg allodynia each responded to only 1 stimulus, the 1 s or 7 s static tactile stimulus (Table 2). As in the radicular pain patients at large, most of those who were mapped closely reported allodynia (16 of 22, 72.7%). This included all 7 of the bilateral patients tested (thus 23 of 29 legs, 79.3%), allowing us to compare the 2 legs in the same individual. The 5 types of stimuli used were delivered at 6 locations—low back, upper thigh, lower thigh, midcalf, dorsal foot, and plantar foot— with 30 tests per dermatome (Fig. 1). Nearly one quarter of these nonnoxious stimuli were reported as painful in the 16 allodynic patients (6.9 ± 6.3 per dermatome in 23 legs). Tactile allodynia was reported on the foot (dorsal and/or plantar) in nearly half of the subjects (7 of 16 patients, 9 of 23 legs). Numbers were similar for cold allodynia (6 of 16 patients, 9 of 23 legs). For any type of allodynia, tactile or cold, the numbers were 8 of 16 patients and 12 of 23 legs. The ranking of allodynia types was similar to the larger group of patients (Tables 2 and 3). In the 7 bilateral patients, allodynia was similar in quality on the 2 legs, with a similar number of allodynia types (P = .18). 3.2. Allodynia on the lower back Radicular patients also reported feeling allodynia on skin of the lower back, although less frequently than on the leg [29.4% of patients tested on the back (10 of 34) vs the 60.8% (45 of 74) who reported leg allodynia; v2 P = .005; Table 4, Fig. 1]. In patients with axial LBP, allodynia on the back occurred at about the same frequency (8 of 15 vs 10 of 34; P = .12). Thus, having pain in the leg did not alter the likelihood of having allodynia on the back. The much larger size of the ventral ramus may explain the predominance of leg allodynia over back allodynia in patients with ventral ramus involvement (ie, radicular patients). Back pain was bilateral in all of the axial LBP patients, and in those who had allodynia on the back, the allodynia was also usually bilateral (7 of 8, 88%). Bilateral back pain was less common in the radicular patients (20 of 34, 56.8%; P = .002), but when it occurred, allodynia on the back was also usually bilateral (8 of 10, 80%). Thus, while allodynia did not always accompany leg or back pain, it only occurred when adjacent underlying deep background pain was present on the same side. There were no instances of mirror allodynia. Allodynia on the leg and back were similar in
Table 4 Proportion of patients tested who had LBP with radiation or axial LBP (LBP without radiation) and reported allodynia on the leg, lower back, or both in the primary dermatome.
a
Location of allodynia
LBP with radiation
Axial LBP
P valuea
Leg Lower back
45/74 (60.8%) 10/34 (29.4%)
2/15 (13.3%) 8/15 (53.3%)
.001 .12
Statistical tests are based on v2 or Fisher’s exact probabilities tests.
Ò
2556
R. Defrin et al. / PAIN 155 (2014) 2551–2559
Table 5 Force threshold for first detecting a light touch stimulus (1 s von Frey) in allodynic skin on the back and midcalf.a Group
n
Back, mN ± SD
Painful leg, mN ± SD
Nonpainful leg, mN ± SD
LBP with radiation Axial LBP
12 10
0.39 ± 0.3 1.3 ± 2.2
1.1 ± 1.6
5 8
Not tested 0.2 ± 0.06
1.05 ± 1.4
1.0 ± 1.7 1.9 ± 2.1 1.4 ± 1.5 1.1 ± 1.3 0.5 ± 0.1 0.5 ± 0.2
CRPS (unilateral) Healthy controls
LBP = low back pain; CRPS = chronic regional pain syndrome. a Corresponding values are provided for the contralateral calf at the mirror-image location on the nonpainful leg. When neither leg was painful (axial LBP and for healthy controls), values for both legs are given under ‘‘nonpainful leg.’’ None of the values differed significantly (P > .05).
quality and intensity (Tables 2 and 3). None of the control subjects reported allodynia on the back. 3.3. Intensity Like other cohorts of radicular pain patients surveyed [26], our patients seldom complained of the presence of tactile or cold allodynia during the interviews or in the MPQ, although other spontaneous sensations such as pins and needles, electrical currents and numbness were mentioned. So as not to draw attention to such symptoms, we did not ask to what extent they contributed to the overall burden of pain. When present, allodynia on the leg and back was rated as moderate (about 4 on the 0–10 VAS; Tables 2 and 3). But even at its worst, LBP patients rated their tactile allodynia as much milder than CRPS patients (P < .001), who uniformly gave VAS scores of 10. Interestingly, the tactile detection threshold in LBP patients, CRPS patients, and healthy controls was similar (Table 5) even though the quality of the sensation reported at threshold stimulus intensity was very different (pain vs touch). This suggests that in all cases, the sensation at threshold was signaled by the same type of low-threshold mechanoreceptive afferents [33,41,46]. Global VAS intensity ratings and MPQ scales measured the patients’ overall pain experience over the past month. The radicular patients uniformly reported more intense and persistent overall pain than those with axial LBP (Table 6), and the intensity and duration of their radicular (leg) pain correlated with that of their back pain (P < .001, Table 7). Likewise, radicular patients with leg allodynia reported more deep back and leg pain than those without on most scales (P < .05, Table 6). A striking exception was the global VAS measure of leg pain intensity, where there was no difference (P = .25). Apparently the development of tactile allodynia, and by inference of central sensitization, has more to do with individual predisposition than with the intensity of the precipitating noxious input. However, when allodynia was present, the global
intensity of leg pain did correlate significantly with the intensity of the allodynia, for 2 subtypes at least: dynamic von Frey and dynamic cotton (P = .01 and .05 respectively, Table 7). Although longitudinal data were not collected, one can infer from the time-series data of patients with progressively longerlasting symptoms that allodynia likely increases over time. Specifically, the prevalence of static tactile allodynia correlated positively with the duration of radicular pain (P = .04) and that of cold allodynia trended in the same direction (P = .06, Table 7). Remarkably, however, there appeared to be a negative relation between intensity (VAS) rating of cold allodynia and the duration of radicular pain (P = .035) and a hint of such a relation also with static tactile allodynia (P = .066, Table 7). This was also true for back pain (P < .01 and .02). Our tentative inference is that as time passed, tactile and cold allodynia faded in intensity just as they were becoming more common. Prospective studies are required to confirm that these static observations indeed represent dynamic processes that evolve over time. The results in hand, however, suggest that allodynia emerges with a delay after the appearance of radicular pain, but with its intensity near maximum at the time of emergence. Subsequently it tends to fade. If this pattern is confirmed by longitudinal data in individual patients followed over time, bedside tracking of the first appearance of tactile and cold allodynia, as well as of progressive changes over time, might provide valuable prognostic information on the likely course of the condition [48]. 4. Discussion The classic signs and symptoms of LBP are well characterized [4,12]. Here we documented the frequent presence of a novel sign, cutaneous tactile allodynia. The sole comparable study [13] tested a single stimulus modality (dynamic mechanical) at a single leg location (foot dorsum). One of 12 patients tested reported marked allodynia. The authors, however, dismissed this case as an outlier. In fact, their result was no different from ours at this test location
Table 6 Comparison of various pain measures in LBP patients with and without pain radiation into the leg.a Pain measure
MPQ-PRI MPQ-NWC MPQ-Pleast MPQ-Pworst Radicular pain (VAS) Back pain (VAS) Radicular pain (duration, y) Back pain duration, y)
LBP with radiation
Axial LBP
Comparison, P value
Allodynia (n = 45)
No allodynia (n = 29)
Allodynia (n = 2)
No allodynia (n = 13)
30.8 ± 14.8⁄ 11.8 ± 5.6⁄ 2 5 6.3 ± 2.9 6.7 ± 1.9⁄⁄ 5.7 ± 5.6⁄⁄ 7.9 ± 7.5
23.2 ± 12.7 9.5 ± 4.8 2 5 5.6 ± 2.7 4.9 ± 3.0 2.7 ± 2.9 10.6 ± 14.7
6.5 ± 0.8 9.1 ± 5.9 2 4 0 4.2 ± 0.3
16.5 ± 13.6 8.1 ± 7.8 1.5 4 0 6.4 ± 1.9
