Impact of nervous system hyperalgesia on pain, disability, and quality of life in patients with knee osteoarthritis: A controlled analysis

Abstract

Objective

Refractory, disabling pain associated with knee osteoarthritis (OA) is usually treated with total knee replacement. However, pain in these patients might be associated with central nervous sensitization rather than peripheral inflammation and injury. We evaluated the presence of hyperalgesia in patients scheduled for a total knee replacement due to knee osteoarthritis with refractory pain, and we assessed the impact of pressure pain threshold measurements (PPT) on pain, disability, and quality of life of these patients.

Methods

Sixty-two female patients were compared with 22 age-matched healthy controls without reported pain for the last year. PPT was measured at the lower extremities subcutaneous dermatomes, over the vastus medialis, adductor longus, rectus femoris, vastus lateralis, tibialis anterior, peroneus longus, iliacus, quadratus lumborum and popliteus muscles and at the supraspinous ligaments from L1–L5, over the L5–S1 and S1–S2 sacral areas and at the pes anserinus bursae and patellar tendon.

Results

Patients with knee OA had significantly lower PPT over all evaluated structures versus healthy control subjects (P < 0.001). Lower PPT values were correlated with higher pain intensity, higher disability scores, and with poorer quality of life, except for the role-emotional and general health status. Combined PPT values over the patellar tendon, at the S2 subcutaneous dermatome and at the adductor longus muscle were the best predictors for visual analog scale and Western Ontario and McMaster Universities Osteoarthritis Index pain scores.

Conclusion

Patients with pain due to osteoarthritis who were scheduled for total knee replacement showed hyperalgesia of nervous system origin that negatively impacted pain, knee functional capacity, and most aspects of quality of life.

INTRODUCTION

Osteoarthritis (OA) is the most common form of arthritis, and is a major cause of pain and disability in the elderly (1), affecting millions of people in the US (2, 3). OA is the fourth most frequently predicted cause of health problems worldwide in women (4). However, because OA has no cure, therapeutic goals are focused on maximizing function and quality of life while controlling pain and minimizing the potentially harmful side effects of medications and therapies (2, 3). Contemporary pain management has shifted from symptom control to management based on the pathophysiologic mechanisms of pain (5). Control of pain in patients with OA remains challenging, and patients with severe OA of the knee who have failed to respond to an extended course of conservative therapeutic modalities are usually scheduled for total knee replacement (1–3).

We are now beginning to have a better understanding of the concepts of peripheral and central sensitization as they relate to knee OA pain (1). Recently, it has been recognized that constant and intense nociceptive sensory information, generated by painful and inflamed deep somatic structures, produces significant neurochemical and metabolic changes, as well as neurologic reorganization within spinal cord segments (6, 7). These changes include an increased excitability of dorsal horn neurons, which in turn produces pain hypersensitivity in a segmental distribution (6). This increased excitability is also known as central sensitization, and both it and peripheral sensitization cause neurons to respond to stimuli in a more intense fashion or to stimuli that they would not normally respond to (5).

Diagnosis of peripheral and central sensitization is very important because spinal cord neurons that normally would only be activated by noxious stimuli can then be activated by normally or typically non-noxious stimuli (allodynia) (8). Together, these neurochemic changes suggest that pain induces, and is partially maintained by, central sensitization (9). Once these complex mechanisms are present, the rationale for treatment approaches should also target central nervous system structures rather than using antiinflammatory agents alone. In fact, it has been speculated that changes in the central nervous system associated with chronic pain might promote peripheral inflammation (10). Importantly, central sensitization may possibly not be attenuated by reversing inflammation in the peripheral tissue (7), and, therefore, refractory pain may persist even after a total knee replacement. Finally, plastic changes in the spinal cord might induce changes in other central structures, such as the limbic and somatosensory cortex, and this may play an important role in the maintenance of chronic pain (11, 12).

Several studies have already investigated the involvement of central pain modulation in OA (1, 13–16). Bajaj et al described deep hyperalgesia in the tibialis anterior muscle of patients with knee OA (13). Creamer et al demonstrated that the injection of local anesthetic in one knee was followed by pain relief in the contralateral noninjected knee (14). Both studies (13, 14) suggested the possible role of the nervous system in the maintenance of chronic pain in patients with knee OA. Assuming that the spinal nerves might be sensitized, we predicted that the dermatomes, myotomes, and sclerotomes innervated by the sensitized spinal nerve would exhibit areas of hyperalgesia. In order to further investigate the phenomenon in the clinical setting, we performed a cross-sectional study assessing superficial and deep hyperalgesia in patients undergoing total knee replacement due to refractory pain associated with OA, and compared them with healthy controls.

Our goal was to evaluate the presence of nervous system sensitization in these subjects. We hypothesized that nervous system hyperalgesia evaluated by decreased pain threshold to pressure both in superficial and deep structures may be a possible cause of pain in this population, and could be correlated with reduction of functional capacity and quality of life in patients with knee OA.

PATIENTS AND METHODS

Patients.

A total of 62 women participated in this study (mean ± SD age 71.10 ± 6.61 years) and data was collected from January to September 2004. We recruited 65 of 321 patients who were on a waiting list for a total knee replacement at the Institute of Orthopedics and Traumatology at the University of Sao Paulo School of Medicine. Women with an established diagnosis of knee OA according to the American College of Rheumatology criteria (17) and Kellgren/Lawrence scale grades of 2–4 (18) were selected. Patients were required to have a pain score ≥4 on a 10-point visual analog scale (VAS) during the week preceding the clinical evaluation. We did not include patients who had the clinical manifestations of OA in other joints, a clinical diagnosis of associated fibromyalgia, a neurologic condition such as stroke or Parkinson's disease, any systemic inflammatory disease, or those who found it impossible to come to the hospital for evaluations. Of the 65 recruited patients, 3 declined to participate in the study. The control group consisted of 22 age-matched female volunteers ages 60 to 85 years (mean ± SD age 68.95 ± 7.40 years) who had no reported pain in the lower back or in the lower extremities for the previous year.

The study was approved by the Ethics Review Committee at the Hospital das Clinicas of the University of Sao Paulo School of Medicine. All eligible patients and healthy volunteers received verbal instructions concerning the study protocol and gave informed written consent to participate in the study prior to enrollment. Patients and volunteers were allowed to withdraw their consent at any time.

Pressure pain threshold measurements.

An experienced rehabilitation physician (HHSK) systematically evaluated superficial and deep hyperalgesia by assessing pressure pain threshold (PPT) measurements using a pressure algometer (Pain Diagnostics, Great Neck, NY) at subcutaneous, myotomal, and sclerotomal structures. First, PPT was measured during the pinch and roll maneuver (as described by Keegan and Garret [19]) at the L1, L2, L3, L4, L5, S1, and S2 dermatome levels to evaluate subcutaneous hyperalgesia (Figure 1).

Figure 1. Pressure pain threshold measurement (in kg/cm²) using a pressure algometer during the pinch and roll maneuver at the L5 dermatome level in a patient diagnosed with knee osteoarthritis.

Myotomal involvement was evaluated by measuring PPT over 9 predefined sites at the vastus medialis, adductor longus, rectus femoris, vastus lateralis, tibialis anterior, peroneus longus, iliacus, quadratus lumborum, and popliteus muscles at classically described painful areas (20). Finally, sclerotomal hyperalgesia was assessed by measuring PPT at the L1–L2, L2–L3, L3–L4, L4–L5 supraspinous ligaments, over the L5–S1 and S1–S2 sacral areas, pes anserinus bursae, and at the patellar tendon. Except for supraspinous ligaments and the L5–S1 and S1–S2 sacral areas (6 sites), all measurements were done bilaterally, with the right side always being evaluated first. Only one trial was performed per site of evaluation per patient. The order of PPT evaluations was standardized and is shown in Figure 2. The PPT is expressed in kg/cm² and its standardized values, validity, and reproducibility have been demonstrated in normal muscles (21). Higher PPT values indicate less severe symptoms.

Figure 2. The anatomic sites for pressure pain threshold evaluations over the muscles, patellar tendon, pes anserinus bursae, and supraspinous ligaments in the anterior, posterior, and lateral views. 1 = vastus medialis muscle; 2 = rectus femoris muscle; 3 = vastus lateralis muscle; 4 = adductor longus muscle; 5 = tibialis anterior muscle; 6 = peroneus longus muscle; 7 = patellar tendon; 8 = pes anserinus bursae; 9 = popliteous muscle; 10 = iliacus muscle; 11 = quadrates lumborum muscle; 12 = supraspinous ligaments, area between L5–S1 and S1–S2.

Pain, disability, and quality of life evaluation.

Intensity of knee pain was assessed on a 10-cm VAS. Anchors on each end were no pain and pain as bad as it could be. Knee function was evaluated using the validated Portuguese version of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (22). The WOMAC questionnaire uses 3 subscale scores: pain, stiffness, and physical activities (difficulty in performing physical activities). Quality of life was assessed using the validated Portuguese version (23) of the Medical Outcomes Study Short Form 36 (SF-36) health survey domains (24). This scale quantifies patients' quality of life in 8 multi-item scales measuring physical functioning, role limitations due to physical health (role-physical), bodily pain, general health perceptions, vitality, social functioning, role limitations due to emotional problems (role-emotional) and mental health. Higher values indicate better quality of life. Questionnaires were completed prior to the PPT tests, on the same day.

Statistical analysis.

Kolmogorov-Smirmov adjustment tested for a normal distribution for all studied variables. PPT values at the subcutaneous dermatomes, muscles, and at the supraspinous ligaments, pes anserinus, and over the patellar tendon were analyzed using a mixed-linear model approach with a restricted maximum likelihood estimation procedure. For the supraspinous ligaments, the analysis included the following covariates: group (knee OA and healthy controls) and site (6 sites of evaluation). For other PPT values, the analysis included the following covariates: subgroup (right and left for healthy controls, affected and nonaffected knee for unilateral OA, more involved and less involved for bilateral OA), and site of evaluation (7 at the dermatomal, 9 at the myotomal, and 2 at the sclerotomal level). Mann-Whitney U test was used to compare all self-reported questionnaire data between the 2 groups. Correlations between VAS, WOMAC scores, SF-36 scores, and the PPT values were evaluated using scatter plots and Spearman's rank correlation coefficients. Stepwise multiple linear regression models analyzed the relationship of VAS, WOMAC, and SF-36 subscales and the PPT measures. Analyses were performed using SPSS software, version 15 (SPSS, Chicago, IL). P values less than 0.05 (2-tailed) were considered significant.

RESULTS

The demographic data of the patients and healthy controls are shown in Table 1. No significant differences between groups were seen in age, marital status, education, race, or occupational activities. Forty-four patients had bilateral knee OA, 12 had knee OA on the right side, and 6 on the left side. In patients with bilateral knee OA, the more symptomatic knee was considered the more affected one. The right side was more involved in 31 and the left side in 13 patients with bilateral involvement. Kolmogorov-Smirnov adjustment test did not demonstrate a normal distribution for any of the studied variables.

Table 1. Clinical and demographic characteristics of knee osteoarthritis (OA) patients and healthy control subjects*

Characteristics	Knee OA (n = 62)	Control (n = 22)	P†
* VAS = visual analog scale; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; SF-36 = Medical Outcomes Study Short Form 36. † By Mann-Whitney U test unless indicated otherwise. ‡ By unpaired t-test. § By chi-square test.
Mean ± SD age, years	71.10 ± 6.61	68.95 ± 7.40	0.38‡
Marital status, n (%)
Married or common-law	32 (51.6)	11 (50.0)	0.91§
Single, divorced, widowed	30 (48.4)	11 (50.0)
Education, n (%)
Illiterate	14 (22.6)	6 (27.3)	0.13§
Elementary	43 (69.3)	11 (50.0)
High school, college, university	5 (8.1)	5 (22.7)
Race, n (%)
White	41 (66.1)	17 (77.3)	0.23§
Black	5 (8.1)	3 (13.6)
Other	16 (25.8)	2 (9.1)
Occupational activity, n (%)
Retired	36 (58.1)	11 (50.0)	0.55§
Housewife	22 (35.5)	8 (36.4)
Working	4 (6.4)	3 (13.6)
Pain duration, mean ± SD months	99.8 ± 75.3
VAS, median (range)	7.9 (4.1–10)	0 (0–0.8)	< 0.001
WOMAC score, median (range)
Pain	55 (20.0–85)	0 (0–0.8)	< 0.001
Stiffness	50 (0–100)	0 (0–0.5)	< 0.001
Physical activity	57.5 (32.3–95.5)	0 (0–0.6)	< 0.001
SF-36 score, median (range)
Physical functioning	15.0 (0–65)	92.5 (40–100)	< 0.001
Role-physical	0 (0–75)	100 (0–100)	< 0.001
Bodily pain	22.0 (0–61)	84.0 (61–100)	< 0.001
General health	71.5 (20–95)	82.0 (5–100)	0.006
Vitality	70.0 (0–95)	87.5 (35–130)	< 0.001
Social functioning	35.51 (12.5–87.5)	100 (37.5–100)	< 0.001
Role-emotional	100 (0–100)	100 (0–100)	0.460
Mental health	72.0 (36–92)	92 (24–100)	0.0002

Dermatomal hyperalgesia.

Our model assessing dermatomal hyperalgesia (25, 26) showed that PPT values were significantly lower in all subcutaneous dermatomes of the knee OA patients when compared with healthy controls (P < 0.001) (Table 2). In the healthy control group, right and left sides presented similar PPT values (P = 0.56), except for a higher PPT measurement at the right side over the L3 subcutaneous area (P = 0.004). In the knee OA group, no difference was observed among diseased knees (more or less affected in bilateral knee OA, and affected in unilateral OA, P = 0.686) or between diseased and the nonaffected knees in unilateral knee OA (P = 0.996).

Table 2. Pressure pain threshold (in kg/cm²) measured at all studied structures*

Variable	Knee OA				Control
	Unilateral noninvolved (n = 18)	Unilateral involved (n = 18)	Bilateral less involved (n = 44)	Bilateral more involved (n = 44)	Right (n = 22)	Left (n = 22)
* Results are expressed as mean/median (95% confidence interval). L = lumbar; S = sacral; VM = vastus medialis muscle; AL = adductor longus muscle; RF = rectus femoris muscle; VL = vastus lateralis muscle; TA = tibialis anterior muscle; PER = peroneus longus muscle; ILIO = iliacus muscle; QL= quadratus lumborum muscle; POP = popliteus muscle; PA = pes anserinus bursae; PT = patella tendon; L1–L2 = L1–L2 supraspinous ligament; L2–L3 = L2–L3 supraspinous ligament; L3–L4 = L3–L4 supraspinous ligament; L4–L5 = L4–L5 supraspinous ligament; L5–S1 = L5–S1 sacral area; S1–S2 = S1–S2 sacral area.
Dermatomal
L1	2.6/2.5 (1.9–3.3)	2.7/2.3 (2.0–3.4)	2.9/2.6 (2.5–3.4)	2.8/2.6 (2.4–3.3)	5.0/4.6 (4.3–5.6)	5.1/5.0 (4.5–5.8)
L2	2.4/2.0 (1.7–3.1)	2.1/2.0 (1.4–2.8)	2.5/2.6 (2.1–3.0)	2.5/2.4 (2.1–3.0)	4.8/4.5 (4.1–5.4)	4.9/4.5 (4.3–5.6)
L3	2.5/2.2 (1.8–3.2)	2.0/2.0 (1.3–2.7)	2.3/2.4 (1.9–2.8)	2.3/2.3 (1.9–2.7)	5.3/4.3 (4.6–5.9)	4.3/3.9 (3.7–4.9)
L4	3.5/3.0 (2.8–4.2)	3.0/3.0 (2.3–3.7)	3.3/3.4 (2.8–3.7)	3.3/3.4 (2.9–3.8)	6.5/6.1 (5.9–7.2)	5.9/5.1 (5.3–6.5)
L5	3.0/2.9 (2.3–3.7)	3.1/2.9 (2.4–3.8)	3.1/3.2 (2.7–3.6)	3.2/3.2 (2.8–3.7)	5.9/5.2 (5.2–6.5)	5.2/4.8 (4.6–5.9)
S1	3.8/3.1 (3.1–4.5)	3.5/3.2 (2.8–4.2)	3.6/3.4 (3.2–4.1)	3.6/3.8 (3.2–4.1)	5.7/5.3 (5.0–6.3)	5.6/5.2 (5.0–6.2)
S2	3.2/2.6 (2.5–3.8)	2.9/2.6 (2.2–3.6)	3.3/3.0 (2.9–3.8)	3.3/3.1 (2.8–3.7)	5.7/5.5 (5.1–6.3)	5.5/5.1 (4.9–6.1)
Myotomal
VM	3.6/3.4 (2.8–4.4)	3.4/3.0 (2.5–4.2)	3.7/3.6 (3.1–4.2)	3.6/3.5 (3.1–4.1)	7.5/6.0 (6.7–8.2)	7.1/6.5 (6.4–7.8)
AL	3.1/2.7 (2.3–3.9)	2.8/2.6 (2.0–3.7)	3.1/2.8 (2.6–3.6)	2.9/2.9 (2.4–3.4)	5.7/5.3 (4.9–6.4)	5.4/5.0 (4.7–6.2)
RF	5.7/5.6 (4.9–6.5)	4.5/4.3 (3.7–5.3)	5.1/5.0 (4.5–5.6)	5.1/5.2 (4.6–5.6)	8.8/8.5 (8.1–9.6)	9.0/9.3 (8.3–9.8)
VL	4.1/3.6 (3.3–4.9)	3.9/3.5 (3.1–4.7)	3.7/3.7 (3.2–4.3)	3.6/3.8 (3.1–4.1)	6.7/5.9 (5.9–7.4)	5.8/5.6 (5.1–6.6)
TA	5.2/4.9 (4.4–6.0)	4.8/4.6 (4.0–5.6)	5.2/5.0 (4.7–5.7)	5.1/4.9 (4.6–5.7)	8.0/7.1 (7.2–8.7)	7.4/6.8 (6.7–8.1)
PER	4.9/4.6 (4.1–5.7)	4.5/4.2 (3.6–5.3)	4.6/4.5 (4.1–5.2)	4.6/4.7 (4.1–5.1)	8.1/8.0 (7.4–8.8)	7.8/7.2 (7.0–8.5)
ILIO	3.7/3.5 (2.9–4.5)	3.4/3.4 (2.5–4.2)	3.5/3.4 (3.0–4.1)	3.4/3.6 (2.9–4.0)	5.1/5.0 (4.4–5.9)	4.9/4.6 (4.2–5.7)
QL	4.8/4.6 (4.0–5.6)	4.2/4.0 (3.4–5.0)	4.5/4.4 (3.9–5.0)	4.5/4.3 (3.9–5.0)	6.3/5.6 (5.5–7.0)	6.1/5.6 (5.4–6.9)
POP	4.1/3.6 (3.2–4.9)	3.7/3.1 (2.9–4.5)	3.4/3.4 (2.9–4.0)	3.4/3.3 (2.8–3.9)	5.8/5.0 (5.1–6.6)	6.0/5.7 (5.3–6.7)
Sclerotomal
PA	3.6/3.3 (2.5–4.6)	2.6/2.7 (1.6–3.6)	3.1/3.2 (2.5–3.8)	3.0/2.9 (2.3–3.6)	6.9/6.0 (6.0–7.8)	5.6/5.1 (4.7–6.6)
PT	5.8/5.5 (4.8–6.8)	5.0/4.8 (4.0–6.0)	5.6/5.4 (5.0–6.2)	5.6/5.6 (5.0–6.3)	11.2/10.5 (10.3–12.1)	10.4/9.8 (9.4–11.3)
L1–L2		5.1/4.8 (4.5–5.6)			9.2/9.1 (8.2–10.1)
L2–L3		4.9/4.7 (4.3–5.5)			9.5/9.8 (8.5–10.5)
L3–L4		4.9/4.6 (4.3–5.5)			9.5/9.5 (8.5–10.4)
L4–L5		5.0/5.1 (4.5–5.6)			9.4/9.0 (8.5–10.4)
L5–S1		5.6/5.5 (5.0–6.2)			9.7/10.2 (8.7–10.7)
S1–S2		5.1/5.0 (4.5–5.7)			9.5/10.1 (8.5–10.5)

Myotomal hyperalgesia.

Analysis of myotomal structures (25, 26) revealed that PPT values were significantly lower in the knee OA patients versus the healthy controls (P < 0.001) (Table 2). In addition to finding similar PPT values between left and right sides in healthy controls (P = 0.511), we found similar PPT values between all diseased knees (P = 0.654) as well as the noninvolved sides in the unilateral OA patients (P = 0.927).

Sclerotomal hyperalgesia.

Our model for assessing sclerotomal hyperalgesia (25, 26), as measured at the supraspinous ligaments, pes anserinus bursae, and patellar tendon, revealed significantly lower PPT measurements at all evaluated structures in the knee OA patients when compared with healthy controls (P < 0.001, Table 2). Of note, within the healthy control group, the right side presented a significantly higher PPT value than the left side at the pes anserinus bursae (P = 0.01). In addition, we found similar PPT values between all diseased knees (P = 0.57) as well as the noninvolved sides in the unilateral knee OA patients (P = 0.127). We did not find any differences among the PPT values at any of the supraspinous ligaments (P = 0.179) in the knee OA group.

Evaluation of pain, disability, and quality of life.

The descriptive statistical data of VAS, WOMAC, and SF-36 scores for both groups are shown in Table 1. WOMAC scores were significantly higher in the knee OA group for all subscales (P < 0.001). SF-36 values for the control group were significantly higher (P < 0.01) than the knee OA group for all subscales except for role-emotional domain (P = 0.46).

We tried to identify the PPT variables that significantly influenced pain intensity (VAS), knee function (WOMAC), and quality of life (SF-36) in OA patients, using a stepwise multiple linear regression model. Initially, all independent variables were tested separately. We found significant correlations in VAS (P < 0.01), in the pain, stiffness, and physical activities subscale scores of the WOMAC questionnaire for all PPT measures (P < 0.05) and for all SF-36 domain values (P < 0.01) except for the role-emotional (P > 0.05) and all PPT values, and also for the general health domains and PPT measurements at the iliacus, quadratus lumborum, and popliteus muscles, at the S1–S2 sacral area and over L1 and S1 dermatomes (P > 0.05). Lower PPT values were correlated with higher pain intensity, higher disability scores, and with poorer quality of life values, except for the role-emotional and general health status.

Stepwise analysis revealed that a combination of 3 variables added gain to the equation. Table 3 shows significant correlations found between PPT values and VAS, WOMAC pain scores, WOMAC physical activity scores, and SF-36 bodily pain scores. Combined PPT values over the patellar tendon, at the S2 subcutaneous dermatome, and at the adductor longus muscle were the best predictors (R² = 0.614) for VAS and the pain subscale of the WOMAC questionnaire (R²=0.608), both correlating inversely (Tables 3 and 4). Combined PPT values over the S2 and L2 subcutaneous dermatomes and at the peroneus longus muscle were the best predictors (R² = 0.611) for the physical activity subscale score of the WOMAC questionnaire, and also correlated inversely (Tables 3 and 4).

Table 3. Correlations between pressure pain threshold values (in kg/cm²) and visual analog scale, WOMAC pain and physical activities subscales, and SF-36 bodily pain score*

Variables	r	P
* WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; SF-36 = Medical Outcomes Study Short Form 36.
Visual analog scale
Patellar tendon	−0.545	< 0.001
S2 subcutaneous level	−0.546	< 0.001
Adductor longus muscle	−0.561	< 0.001
WOMAC pain
Patellar tendon	−0.589	< 0.001
Adductor longus muscle	−0.540	< 0.001
S2 subcutaneous level	−0.601	< 0.001
WOMAC physical activity
S2 subcutaneous level	−0.509	< 0.001
Peroneus longus muscle	−0.571	< 0.001
L2 subcutaneous level	−0.550	< 0.001
SF-36 bodily pain
Adductor longus muscle	0.569	< 0.001
L1–L2 supraspinous ligament	0.534	< 0.001
L2 subcutaneous level	0.606	< 0.001

Table 4. Stepwise multiple linear regression models for VAS, WOMAC pain subscale, WOMAC physical activity subscale, and SF-36 bodily pain subscale and predicted pressure pain threshold values (in kg/cm²)*

Stepwise multiple linear regression model	R^2
* VAS = visual analog scale; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; SF-36 = Medical Outcomes Study Short Form 36; PT = patellar tendon; S2 = S2 subcutaneous level; AL = adductor longus muscle; PER = peroneus longus muscle; L2 = L2 subcutaneous level; L1–L2 = L1–L2 supraspinous ligament.
VAS = 12.41 − 0.39(PT) − 0.52(S2) − 0.55(AL)	0.614
WOMAC pain = 91.82 − 3.11(PT) − 4.28(AL) − 3.5(S2)	0.608
WOMAC physical activity = 98.33 − 3.85(S2) − 4.37(PER) − 5.84(L2)	0.611
SF-36 bodily pain = −8.87 + 6.12(AL) + 2.96(L1–L2) + 3.57(L2)	0.618

Lower PPT values at the adductor longus muscle, L1–L2 supraspinous ligament and at the subcutaneous L2 area were associated with poorer bodily pain domain of the SF-36 (R² = 0.618) (Tables 3 and 4). Combined PPT values at the adductor longus muscle, at the subcutaneous S2 dermatome, and at the L1–L2 supraspinous ligament were associated with lower SF-36 total score (R² = 0.505) (Tables 3 and 4).

DISCUSSION

Similar to the results observed in our pilot study (27), this study demonstrated that patients with moderate to severe persistent knee pain and disability, which was not relieved by an extended course of nonsurgical treatment, and therefore scheduled for a total knee replacement, had significantly lower pressure pain thresholds (P < 0.001) versus healthy controls (27). We demonstrated that the differential PPT threshold between knee OA patients and healthy controls was constant throughout sites of assessment at the dermatomal, myotomal, and sclerotomal structures. Correlation analysis and multiple linear regression scatterplots determined the relationship between VAS, WOMAC, and SF-36 subscales, and PPT measures (Tables 3 and 4). We found significant correlations between all these measurements (P < 0.01), except for 2 domains of the SF-36: role-emotional and general health. Lower PPT values were correlated with higher pain intensity, higher disability scores, and poorer quality of life.

One strength of this study was its focus on an approach to clinically identify nervous system hyperalgesia in patients with disabling knee OA pain, an issue that has not been addressed sufficiently in the past. A few other studies have directly assessed central sensitization in OA (1, 13–16). Cutaneous and deep hyperalgesia have been demonstrated in the forearm of patients with thumb-base OA (15), and deep hyperalgesia has been shown in the tibialis anterior muscle of patients with knee OA (13). Our study demonstrated a generalized state of hyperalgesia, both in superficial and deep structures, in knee OA patients when compared with healthy controls. This suggests that the peripheral and central nervous system might be involved in the maintenance of the chronic pain state. The clear understanding of the mechanisms involved in how knee OA pain is generated, and how the sensory information is processed from peripheral receptors to cerebral cortex, might provide useful insights that can lead to clinical benefits in the future. Initially, hypersensitivity is found at the site of damage; however when the disease process is not controlled, such as in patients with OA and refractory pain, the central nervous system undergoes plastic changes that are responsible for sustaining chronic pain. It then becomes independent from the peripheral pathologic process.

Usually, repeated stimulation causes most sensory organs to become fatigued and less responsive (28). High-threshold polymodal C fibers involved in nociception, however, show the opposite response (29). In fact, with repeated nociceptive stimulation, nerve endings undergo changes that result in enhanced sensitivity, lowered threshold to stimulation, and prolonged and enhanced response to the stimulation, also known as after-discharge. This phenomenon is called sensitization and it is responsible for sustained pain, tenderness, and segmental and suprasegmental reflex responses (28).

Our findings demonstrated that centrally induced neuroplastic changes measured by a decreased PPT over superficial and deep structures occurred also in sites distant from the knee area. We showed that PPT values were significantly lower in all evaluated structures. It should be underscored that sensitization associated with chronic pain is observed in all levels of the nervous system, from peripheral structures (receptors and nerves) to central structures (spinal cord and brain). In fact, spinal segmental sensitization is a hyperactive state of the spinal cord caused by repeated stimulation of nociceptive receptors from impulses sent by sensitized damaged tissue to the dorsal horn neurons (central nervous system sensitization). The mechanisms of spinal segmental sensitization include neuron hypertrophy and up-regulation of excitatory neurons and of prohyperalgesic peptides, and neurotransmitters at the dorsal horn of the spinal cord. This results in a mismatch of inflammation and pain, as pain does not indicate worsening of inflammation and vice versa.

Knowledge of the segmental distribution of sensory nerve fibers is important in managing patients with pain (28). In humans, the innervation of the skin, muscles, and deep structures is determined embryologically at an early stage of fetal development, and there is little inter-subject variability (30). Each segment of the spinal cord and its corresponding spinal nerves have a consistent segmental relationship that allows the clinician to ascertain the probable spinal level of dysfunction based on the pattern of dermatomal, myotomal, and sclerotomal hyperalgesia (25, 26). We decided to evaluate PPT over anatomic structures innervated by different segmental spinal nerves and divide them into 3 different categories based on the segmental innervations: dermatomal, myotomal, and sclerotomal (28). For the myotomes, we chose to evaluate the PPT over muscles innervated by different spinal nerves. We selected vastus medialis, vastus lateralis, rectus femoris, and adductor longus as muscles innervated by the L2–L3–L4 spinal nerve roots; tibialis anterior and deep peroneus longus by L5–S1 spinal nerves; iliacus for L1–L2; quadrates lumborum for L2–L3; and popliteus for S1–S2 spinal nerve roots. For the dermatomes we followed those described by Keegan and Garrett (19), and the ones described by Bonica (28) for the sclerotomes.

Our data suggested that hyperalgesia over some structures presented a stronger correlation to pain, disability, and quality of life. Of note was that the combined PPT values over the patella tendon, at S2 subcutaneous dermatome and at the adductor longus muscle were the best predictors for VAS and WOMAC pain. A linear increase in VAS (R² = 0.614) and in WOMAC pain (R² = 0.608) was observed for the lower values of these 3 combined independent variables, indicating that pain may be influenced by involvement of central sensitization at L3, L4, and S2 spinal segments. Similarly, lower PPT values at the adductor longus muscle, L1–L2 supraspinous ligament, and at the L2 dermatome were associated with poorer SF-36 bodily pain domain. These findings are important due to strong biologic rationale, since knees are predominately innervated by L2, L3, L4, and S2 rather than S1 spinal nerves.

Another possible explanation of our findings is that lower PPT values over a spontaneously painful human skeletal muscle presents significantly elevated levels of substance P, calcitonin gene-related peptide, bradykinin, tumor necrosis factor α, interleukin-1β, serotonin, and norepinephrine versus nonpainful, healthy subjects (31). The concentration of selected inflammatory mediators, neuropeptides, cytokines and catecholamines also differ quantitatively from a remote, uninvolved site (32).

It is also important to note that central plastic changes occur not only in the spinal cord, but also in other structures of the central nervous system. For instance, a previous study evaluated brain areas involved in processing pain in a group of 12 patients with OA using FDG. The results showed that arthritic pain was associated with increased activity in the cingulate cortex, the thalamus, and the amygdale; areas involved with the processing of the emotional aspects of pain. Indeed, the authors suggested in this study that new treatments for pain in arthritis should target central structures (33).

Our study has some limitations. First, we did not measure PPT over thoracic and cervical innervated areas, and therefore cannot evaluate whether central sensitization is systemic or concentrated in areas near or related to the knees. However, our results showed that even when OA was unilateral, both extremities were equally affected in terms of hyperalgesia. Further studies should explore hyperalgesia in distant areas. Second, we did not evaluate changes in sensitization in central structures such as cortical brain areas; therefore, we cannot rule out that hyperalgesia is neither associated with nor a result of brain changes. Finally, the analyses reported in this study were exploratory and hypothesis-generating rather than confirmatory. Further studies are necessary to confirm our findings and fully investigate the mechanisms of the nervous system that enhance patients' reported pain. These insights may then encourage further studies to assess new therapeutic approaches to control pain in knee OA.

Patients should be evaluated for nervous system sensitization before surgical procedures are prescribed to rule out these important and easily-detectable clinical findings. PPT values should be performed at least at the S2 dermatomal level, at the adductor longus muscle, and at the patellar tendon structures. These, when combined showed an inverse correlation with and were the best predictors for VAS and WOMAC pain. Measurements should be performed on both sides for patients with knee OA, because decreased PPT on the nonaffected or on the less-affected side is further evidence for possible nervous system mediated hyperalgesia.

Once central nervous system sensitization is present, the rationale for pain treatment should also target the nervous system structures, rather than using antiinflammatory agents alone. The understanding and clinical identification of nervous system sensitization, and its potential for modulation, may provide exciting and innovative cost effective therapeutic tools to control pain, reduce disability, and improve quality of life in knee OA patients.

AUTHOR CONTRIBUTIONS

Dr. Marta Imamura had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Marta Imamura, Satiko Imamura, Hsing, Camanho.

Acquisition of data. Kaziyama, Targino, de Souza, Cutait.

Analysis and interpretation of data. Marta Imamura, Hsing, Fregni, Camanho.

Manuscript preparation. Marta Imamura, Fregni.

Statistical analysis. Fregni.

Acknowledgements

The authors would like to thank David A. Cassius, MD for his assistance in reviewing the manuscript.