This page is an English translation of the Korean original. If there is any discrepancy, the Korean version prevails.

Diffuse Eosinophilic Fasciitis (Shulman Syndrome)미만성 근막염 (미만성근막염)

📋 Contents and key points
ItemDetails
A. Classification codesICD-10: M35.4 | Special Copayment Reduction: V139 (rare disease Special Copayment Reduction)
B. SynonymsDiffuse (eosinophilic) fasciitis, Shulman syndrome, Eosinophilic fasciitis (EF)
C. Adult onsetLikelihood of first onset in adulthood high — typically develops in adults in their 40s–60s and is rarely reported in children. Onset in adults after strenuous physical activity is the typical pattern.
1. History1974 Shulman first described eosinophilic fasciitis → named 'Shulman syndrome' → 1975 Rodnan distinguished it from eosinophilic fasciitis. 1980s increase in cases occurring after strenuous exercise.
2. CausesCause unknown. Associated with onset after strenuous exercise (excessive physical exertion). Associations reported with drugs (benzodiazepines, contaminated L-tryptophan, etc.), infection (Borrelia), and hematologic disorders (thrombocytopenia, aplastic anemia).
3. GeneticsGenetic predisposition unknown. Few HLA association studies. Possible association with a family history of autoimmune disease. No clear genetic variant established.
4. General epidemiologyRare disease. On the order of several hundred case reports worldwide. Incidence and prevalence not established. Somewhat more common in adult (40s–60s) men. Also reported in children. Strenuous physical activity is a major trigger.
5. PathophysiologyTGF-β- and IL-5-driven activation of eosinophils and fibroblasts → increased fascial thickness, collagen deposition. Immune-mediated fascial fibrosis mechanism. Infiltration of CD8+ T cells and macrophages also involved.
6. SymptomsFirm swelling of the skin and underlying tissue of the limbs (especially upper and lower limbs), pain, limited range of motion, orange-peel skin texture. Morning stiffness. Raynaud phenomenon rare. Mild systemic symptoms.
7. SignsSkin depression finding (groove sign: depression along veins on pressure), induration and thickening of the limb skin. Eosinophilia (about 60–70%). Elevated ESR and CRP. EMG normal or nonspecific.
8. ScreeningNo specific screening test. The combination of limb skin induration + eosinophilia is a screening indicator. MRI (fascial enhancement) and biopsy (deep wedge biopsy) are central to diagnosis.
9. DiagnosisPinal-Fernandez et al. criteria: major criteria (limb induration and swelling, histology of eosinophilic fascial infiltration), minor criteria (peripheral eosinophilia, fascial abnormality on MRI, preceding strenuous exercise). Wedge biopsy of skin, fascia, and muscle is confirmatory.
10. TreatmentPrednisolone (0.5–1 mg/kg/day) is first-line therapy. Combination with hydroxychloroquine or methotrexate. IVIG and cyclosporine reported. Physical therapy essential. Separate treatment when a hematologic disorder is present.
11. PrognosisSome achieve complete remission with steroid treatment. Skin induration persists in about 30–50%. Joint contractures may remain. Poor prognosis with an associated hematologic disorder. May progress to fibrosing diseases such as morphea.
12. PreventionPrimary prevention not possible. Early evaluation for limb swelling or induration after strenuous exercise. Management of adverse effects during steroid treatment. Physical therapy to prevent contractures.
13. ReferencesPinal-Fernandez I et al. … and 9 other references

1. History

Diffuse eosinophilic fasciitis was first described in 1974 by the American rheumatologist Lester E. Shulman and is also called 'Shulman syndrome' after him. Shulman reported a syndrome in two patients in whom the skin of the limbs became hard and swollen after strenuous physical activity, and described marked eosinophilic infiltration and fibrosis of the fascia on skin biopsy.

The following year, in 1975, the American rheumatologist Gerald Rodnan reported more patients and described the clinical and histologic features in detail, establishing the term 'eosinophilic fasciitis'. Rodnan emphasized that this disease is a distinct entity separate from systemic sclerosis. As key distinguishing features, he cited the absence of Raynaud phenomenon, no visceral involvement, eosinophilic infiltration on biopsy, and response to steroids.

In the 1980s, cases of eosinophilic fasciitis associated with benzodiazepines and contaminated L-tryptophan supplements (eosinophilia-myalgia syndrome, EMS) were reported, raising awareness of possible drug induction. In 1989, there was a large EMS outbreak (related to contaminated L-tryptophan) in the United States, which gave new impetus to eosinophilic fasciitis research.

After MRI was introduced into clinical practice in the 1990s, fascia-specific signal abnormalities in eosinophilic fasciitis were confirmed by imaging, and skin-fascia biopsy was standardized. In 2014, Pinal-Fernandez et al. systematically reviewed existing cases and proposed diagnostic criteria, laying the foundation for clinical research. Eosinophilic fasciitis is currently an extremely rare disease with only a few hundred cases reported worldwide, and because there are no large randomized controlled trials, the evidence for treatment depends mainly on case reports and small cohorts.

In Korea, reports have been sporadic, with several dozen case reports in the journals of the Korean College of Rheumatology and the Korean Dermatological Association. It is designated as eligible for the rare disease Special Copayment Reduction (V139), and educating physicians for early diagnosis and appropriate treatment is an important task.

2. Causes

The exact cause of diffuse eosinophilic fasciitis is unknown. Taking together the associated factors identified to date, it is understood that in individuals with an immunologic predisposition, various triggers set off an abnormal immune response and fibroblast activation, leading to fascial fibrosis.

Physical overload: In about 30–46% of all patients, unusually strenuous physical exercise or excessive manual labor is reported in the weeks before onset. Typical triggering activities include long-distance marathons, excessive mountain climbing, and heavy labor. A mechanism in which repeated mechanical stimulation of the fascia triggers an immune response is presumed, but the exact mechanism is unknown.

Drugs and toxins: Some drugs, such as benzodiazepines (clonazepam), contaminated L-tryptophan, carbidopa, and simvastatin, have been reported to cause a syndrome similar to eosinophilic fasciitis. L-tryptophan-related EMS (eosinophilia-myalgia syndrome) was found to be caused by contaminants in specific manufacturing batches of L-tryptophan. There are also a few reports of an association with silicone breast implants.

Infection: An association with Borrelia burgdorferi (the causative organism of Lyme disease) has been reported in European cases. There are also cases occurring after parvovirus B19 and other viral infections. However, no single causative infectious agent has been established, and infection is presumed to act as a trigger that induces the immune response.

Association with hematologic disorders: There are reports that hematologic disorders (aplastic anemia, thrombocytopenia, hemolytic anemia, leukemia, lymphoma) are present in about 10–20% of patients with eosinophilic fasciitis, so hematologic evaluation at diagnosis is important. The mechanism of this association is unknown, but it has been suggested that abnormal immune cell activation in the bone marrow may promote fascial inflammation.

3. Genetics

The genetic basis of diffuse eosinophilic fasciitis has not yet been sufficiently elucidated. Because the number of patients is very small, large-scale genetic studies are difficult, and most genetic information is based on small case analyses and individual reports.

Some HLA association studies have suggested associations with specific HLA alleles, but no reproducible results have been established. A family history of autoimmune diseases (rheumatoid arthritis, Sjögren syndrome, Hashimoto thyroiditis, etc.) has been reported in some patients with eosinophilic fasciitis, raising the possibility of an autoimmune genetic predisposition. In addition, variants in genes related to regulatory T-cell function (FOXP3, etc.) and in TGF-β signaling pathway genes are known predispositions to fibrosing diseases in general, and abnormal activation of this pathway is also involved as a key mechanism in eosinophilic fasciitis.

From an immunologic perspective, genetic variants involved in eosinophil activation and hypersecretion of cytokines such as TGF-β and IL-5 may contribute to susceptibility to this disease. No genome-wide association studies (GWAS) or whole-exome sequencing studies have yet been performed, so it remains an area in which the search for rare genetic variants is needed. In current clinical practice, genetic testing is not used for diagnosis or prognostic assessment.

4. General epidemiology

Diffuse eosinophilic fasciitis is an extremely rare disease worldwide, and its exact incidence and prevalence have not been established. Since Shulman's first report in 1974, an estimated 300–500 cases had been reported in the literature worldwide by 2024. This extremely small number of cases makes it difficult to produce reliable epidemiologic statistics.

Age at onset is broad, but it occurs mainly in adults aged 30–70 years, with a median in the 40s–50s. Pediatric-onset cases have been reported but account for less than 10% of the total. The sex distribution differs between studies, but most cohorts report a male predominance (about 55–65%) or an equal sex distribution. Geographically, it is reported more in Europe and North America, but this may reflect differences in access to care and awareness.

There are no incidence data for Korea, but estimates based on the number of patients registered for the rare disease Special Copayment Reduction (V139) suggest it is very rare, at the level of several dozen patients. There are sporadic case reports in Korean rheumatology and dermatology journals, and the fact that diagnostic delay averages 1–3 years because of the difficulty of early diagnosis suggests a similar challenge in Korea. Because it is rare worldwide, there are no randomized controlled trials, and the evidence for treatment depends on observational studies and expert consensus.

5. Pathophysiology

The central role in the pathogenesis of diffuse eosinophilic fasciitis is played by eosinophil and fibroblast activation driven by TGF-β (transforming growth factor β) and IL-5. Immune-mediated fascial fibrosis is understood as the core pathophysiology.

Eosinophil activation: When an initial inflammatory response is triggered by inciting factors (strenuous exercise, drugs, infection, etc.), peripheral blood eosinophils are recruited into tissues by IL-5 and eotaxin. Tissue eosinophils become activated and secrete toxic proteins such as ECP (eosinophil cationic protein), MBP (major basic protein), and reactive oxygen species. These eosinophil secretory products directly activate fascial fibroblasts and stimulate TGF-β secretion.

TGF-β-driven fibrosis: TGF-β differentiates fibroblasts into myofibroblasts and induces hypersecretion of collagen (particularly types I and III). At the same time, it shifts the balance between collagen-degrading enzymes (MMPs) and their inhibitors (TIMPs) toward fibrosis. As a result, irreversible collagen deposition and thickening occur in the fascia, which form the anatomic basis of skin induration, swelling, and limited range of motion.

Role of T cells and macrophages: In addition to eosinophils, the cells infiltrating the fascia include CD4+/CD8+ T cells, macrophages, and plasma cells. CD8+ T cells can directly damage fascial cells by cytotoxic mechanisms, and macrophages secrete TGF-β and IL-1β to promote fibrosis. Reduced regulatory T-cell (Treg) function is presumed to contribute to the failure to suppress excessive inflammatory and fibrotic responses.

Fascial specificity: Why lesions in eosinophilic fasciitis concentrate in the deep fascia rather than the dermis or subcutaneous fat has not yet been fully explained. One hypothesis is that the fascia is an anatomic compartment with limited access for immune cells, so that once inflammation is established, local immunosuppressive mechanisms are insufficient and fibrosis can progress rapidly.

6. Symptoms

The main symptoms of diffuse eosinophilic fasciitis are induration, swelling, and pain of the skin and subcutaneous tissue of the limbs, and the resulting limitation of joint range of motion. Symptoms develop subacutely over days to weeks and are often initially mistaken for simple muscle pain or injury.

Limb skin induration and swelling: The most common and characteristic symptom. Induration is thickening of the skin until it becomes hard like wood; it usually begins in the proximal limbs (forearm, upper arm, thigh, lower leg) and may spread distally. The skin surface shows an orange peel (peau d'orange) texture, and a characteristic feature is induration adherent to the underlying tissue so that the skin cannot be pinched. Early on, pitting edema with reddish skin may appear.

Pain and stiffness: Pain and stiffness of the affected areas are present. They are particularly severe on waking in the morning (morning stiffness) or after prolonged static posture. Flexion of the elbows, knees, and other joints is limited, making daily activities (rolling up sleeves, raising the arms, bending the knees) difficult.

Joint contractures: In a chronic course, flexion contractures due to fascial fibrosis may develop around joints. The wrists, fingers, elbows, and knees are particularly affected, and the later treatment is started, the more irreversible contractures become fixed.

Systemic symptoms: Mild fever, fatigue, weight loss, and muscle pain may be present, but they are milder than in systemic sclerosis. Raynaud phenomenon is very rare in eosinophilic fasciitis (<5%), and its presence requires differentiation from systemic sclerosis. Visceral involvement (pulmonary fibrosis, esophageal dysmotility, renal involvement) is almost absent.

Hematologic abnormalities: Some patients have hematologic disorders such as aplastic anemia, thrombocytopenia, and autoimmune hemolytic anemia, in which case fatigue, bleeding tendency, and pallor may appear.

7. Signs

The most characteristic physical sign of diffuse eosinophilic fasciitis is the 'groove sign'. When the upper limb is raised above the head, the skin is depressed along the course of the superficial veins; this reflects thickening of the surrounding fascia, with the skin over the veins tethered to the soft tissue. The groove sign is a relatively specific clinical sign of eosinophilic fasciitis.

Skin findings: The skin of the affected limbs is hard like wood and adherent to the underlying tissue so that it cannot be pinched. The skin surface shows an orange-peel texture, and its color may turn brown. Unlike systemic sclerosis, the fingertips (acrosclerosis) are usually not involved, and the absence of finger skin induration (sclerodactyly) is an important point distinguishing it from scleroderma.

Joint range of motion: Passive range of motion is limited in joints with flexion contractures. Muscle strength itself is preserved, but function is reduced because of pain and contractures.

Laboratory findings: Peripheral blood eosinophilia appears in about 60–70% (absolute eosinophil count >500/μL, markedly elevated in some). Elevated ESR and CRP, hypergammaglobulinemia, and elevated IgG are observed. Serum aldolase and CK are usually normal or mildly elevated. RF and ANA may be positive at low titers in some, but high-titer positivity suggests another connective tissue disease. Complete blood count may show anemia and thrombocytopenia (when a hematologic disorder is present).

MRI findings: T2 hyperintensity of the fascia (fascial enhancement with gadolinium) is characteristic, and abnormalities at the subcutaneous tissue-fascia-muscle interfaces are observed. MRI is useful for selecting the biopsy site and assessing the extent of lesions.

Histologic findings: The key to diagnosis is a deep wedge biopsy including skin, subcutaneous fat, fascia, and muscle. Infiltration of the fascia by eosinophils, lymphocytes, plasma cells, and macrophages, with thickening and fibrosis of collagen fibers, is characteristic. Changes in the dermis and subcutaneous fat are mild, which is why the diagnosis is missed on superficial skin biopsy.

8. Screening

Because diffuse eosinophilic fasciitis is a very rare disease, there is no population screening. However, it is important to suspect the disease early and evaluate it systematically when the following combination of clinical features is present.

Key clinical screening indicators: Eosinophilic fasciitis should be actively suspected when limb skin induration and swelling appear within weeks after strenuous physical activity in an adult aged 30–70 years, accompanied by peripheral blood eosinophilia. In particular, the combination of 'skin induration + eosinophilia' without Raynaud phenomenon, without finger skin induration, and without symptoms of visceral involvement is the key to diagnostic screening.

Basic blood tests: Measurement of the absolute eosinophil count on a complete blood count (CBC with differential) is essential. ESR, CRP, immunoglobulins (IgG, IgE), RF, ANA, anti-Scl-70 (anti-topoisomerase I), anticentromere antibodies, CK, aldolase, liver function tests, and serum protein electrophoresis are performed to evaluate the differential diagnosis and comorbidities.

MRI: Limb MRI (fat-suppressed T2 and gadolinium contrast enhancement) is essential for noninvasively confirming fascial abnormalities and selecting the optimal biopsy site. The biopsy is planned by correlating areas with the groove sign with areas abnormal on MRI.

Tissue biopsy: For confirmation, a deep wedge biopsy including skin, subcutaneous fat, fascia, and muscle from an affected area is essential. Superficial skin biopsy (punch biopsy) alone is insufficient for diagnosis. Using MRI or ultrasound guidance during biopsy ensures that fascia is included.

Hematologic evaluation: When eosinophilic fasciitis is diagnosed, a peripheral blood smear and the need for bone marrow examination are reviewed to exclude associated hematologic disorders (aplastic anemia, thrombocytopenia, lymphoma, etc.). Consultation with a hematologist is needed especially when cytopenias are present or the response to treatment is poor.

9. Diagnosis

Diagnosis of diffuse eosinophilic fasciitis relies on a synthesis of clinical, laboratory, imaging, and histologic findings. There is currently no single internationally accepted set of classification criteria, but the criteria proposed by Pinal-Fernandez et al. in 2014 are currently the most widely referenced.

Pinal-Fernandez diagnostic criteria (2014): Diagnosis is made when 2 major criteria and at least 2 minor criteria are met. Major criteria: ① symmetric or asymmetric skin induration/thickening and swelling of the limbs, ② eosinophilic infiltration and fibrosis of the fascia on a biopsy including skin, subcutaneous fat, fascia, and muscle. Minor criteria: ① peripheral blood eosinophilia, ② abnormal fascial signal on MRI (T2 hyperintensity or contrast enhancement), ③ history of preceding strenuous physical activity, ④ response to steroid treatment, ⑤ exclusion of other connective tissue diseases.

Differential diagnosis: The following must be excluded: systemic sclerosis (distinguished by Raynaud phenomenon, finger induration, visceral involvement, and positive anti-Scl-70 and anticentromere antibodies), localized scleroderma (morphea), eosinophilia-myalgia syndrome (EMS, history of L-tryptophan or toxin exposure), myositis/dermatomyositis (elevated CK, abnormal EMG, inflammatory findings on muscle biopsy), deep vein thrombosis, lymphedema, reactive eosinophilia (parasitic infection, drug hypersensitivity, hematologic malignancy), and connective tissue diseases other than systemic sclerosis.

Autoantibody testing: High-titer positivity for anti-Scl-70 antibodies, anticentromere antibodies, ANA, and RF suggests systemic sclerosis and other connective tissue diseases. In eosinophilic fasciitis, these are negative or at low titers.

Electrophysiologic testing: Electromyography (EMG) and nerve conduction studies are usually normal or nonspecific. The absence of spontaneous fibrillation potentials and abnormal insertional activity seen in myositis helps differentiate it from myositis.

10. Treatment

It should be recognized that the evidence for treatment of diffuse eosinophilic fasciitis is based mainly on case reports and small retrospective cohort studies, and there is no evidence from randomized controlled trials. Corticosteroids are currently first-line treatment, and immunosuppressants and other treatments are used in refractory or relapsing patients.

Corticosteroids: Prednisolone is started at 0.5–1 mg/kg/day (usually 30–60 mg/day) and tapered slowly as symptoms and laboratory findings improve. About 70–80% of patients respond to steroids, with higher response rates particularly when treated early. Eosinophil counts fall rapidly with steroids, but improvement of skin induration takes several months. Steroid dose is adjusted based on clinical response, eosinophil count, and ESR/CRP.

Hydroxychloroquine: Used in combination with steroids or as a maintenance agent during steroid tapering. 200–400 mg/day orally. Some antifibrotic and immunomodulatory effects have been reported. Ophthalmologic examination (monitoring for retinal toxicity) is needed.

Methotrexate (MTX): In relapsing or steroid-refractory patients, MTX 7.5–15 mg per week is added. It is reported to have a steroid-sparing effect and to help suppress fibrosis. Folic acid supplementation is essential.

Other immunosuppressants: Cyclosporine A, azathioprine, and mycophenolate mofetil are used in refractory cases. Intravenous immunoglobulin (IVIG) has been reported to be effective in some severe eosinophilic fasciitis. There are also cases of use of cyclophosphamide, TNF-α inhibitors (infliximab), and JAK inhibitors (tofacitinib), but the level of evidence is low.

Physical therapy: Physical therapy is a key component of treatment for preventing and recovering from skin induration and joint contractures. Range-of-motion exercises, stretching, ultrasound therapy, and the like are used. Combining drug therapy and physical therapy is essential for functional recovery.

Treatment of associated hematologic disorders: When aplastic anemia, thrombocytopenia, lymphoma, or the like is present, standard treatment for it is needed separately, and hematologic treatment sometimes also helps the eosinophilic fasciitis itself.

11. Prognosis

The prognosis of diffuse eosinophilic fasciitis varies from patient to patient. Complete remission is possible in patients who receive early diagnosis and aggressive steroid treatment, but in some patients skin induration and joint contractures persist.

Remission rate: Complete or near-complete remission is reported in about 50–60% of patients treated with steroids, while partial remission or residual skin induration persists in about 30–40%. The earlier treatment begins, the higher the complete remission rate. When diagnosis is delayed (more than 1–2 years), irreversible contractures become fixed and the prognosis is poor.

Relapse: Relapse occurs in about 25–30% during steroid tapering. With relapse, the steroid dose is increased again and addition of a steroid-sparing agent such as methotrexate is considered.

Prognosis with associated hematologic disorders: When hematologic disorders such as aplastic anemia or lymphoma are present, the prognosis is markedly poor, and the hematologic disorder itself may be the cause of death. In this case, consultation with a hematologist-oncologist is essential.

Long-term complications: If fibrosis persists, joint contractures, muscle atrophy, and chronic pain may remain. Rare cases progressing to morphea or localized scleroderma are reported. Quality of life is significantly reduced by pain, functional limitation, and cosmetic skin changes.

12. Prevention

Primary prevention of diffuse eosinophilic fasciitis is not possible. Prevention strategies focus on early recognition, physical therapy to prevent contractures, and prevention of treatment adverse effects.

Education for early diagnosis: It is important to raise awareness among internists, rheumatologists, and dermatologists so that they suspect this disease early in patients showing the combination of limb skin induration and eosinophilia after strenuous physical activity. Diagnostic delay leads to irreversible contractures and worsens the prognosis.

Prevention of contractures through physical therapy: Physical therapy is started together with treatment to maintain joint range of motion as much as possible. Passive and active joint exercises, aquatic therapy, and limb stretching are performed regularly.

Drug-related prevention: If limb induration or swelling develops in a patient taking a drug associated with eosinophilic fasciitis (benzodiazepines, etc.), the patient should consult a medical professional immediately. Quality control of L-tryptophan supplements is important for preventing EMS.

Management of steroid adverse effects: To prevent adverse effects of long-term steroid treatment (osteoporosis, diabetes, infection, etc.), calcium and vitamin D supplementation, blood glucose and blood pressure monitoring, and up-to-date vaccinations are provided.

13. References

  1. Pinal-Fernandez I, Selva-O'Callaghan A, Grau JM. Diagnosis and classification of eosinophilic fasciitis. Autoimmun Rev. 2014;13(4-5):379-382. DOI: 10.1016/j.autrev.2014.01.019. PMID: 24424181.
  2. Shulman LE. Diffuse fasciitis with hypergammaglobulinemia and eosinophilia: a new syndrome? J Rheumatol. 1984;11(4):569-570. PMID: 6490434.
  3. Lebeaux D, Sene D. Eosinophilic fasciitis (Shulman disease). Best Pract Res Clin Rheumatol. 2012;26(4):449-458. DOI: 10.1016/j.berh.2012.07.001. PMID: 23040358.
  4. Haiduc VF, Erkan D, Kirou K, Bhatt DL. Anti-citrullinated protein antibodies and eosinophilic fasciitis. Clin Exp Rheumatol. 2016;34(1):103-107. PMID: 26641831.
  5. French LE, Shapiro M, Junkins-Hopkins JM, et al. Eosinophilic fasciitis and eosinophilic cellulitis in a patient with abnormal T lymphocytes: increased production of interleukin 5 and relationship to Wells' syndrome. J Am Acad Dermatol. 2003;48(3):465-469. DOI: 10.1067/mjd.2003.144. PMID: 12637930.
  6. Varga J, Griffin R, Newman JH, Jimenez SA. Eosinophilic fasciitis is clinically distinguishable from the eosinophilia-myalgia syndrome and is not associated with L-tryptophan use. J Rheumatol. 1991;18(2):259-263. PMID: 2023205.
  7. Jinnin M, Ihn H, Yamane K, et al. Serum levels of tissue inhibitor of metalloproteinase-1 and 2 in patients with eosinophilic fasciitis. Br J Dermatol. 2004;151(5):1140-1142. PMID: 15541107.
  8. Bonnekoh H, Scheffel J, Magerl M, et al. Eosinophilic fasciitis with concomitant hematologic malignancy. Acta Derm Venereol. 2018;98(1):131-132. DOI: 10.2340/00015555-2783. PMID: 28749530.
  9. Kim YJ, Bae JP, Kim JY, et al. Eosinophilic fasciitis: a case series of Korean patients. J Dermatol. 2020;47(6):664-667. DOI: 10.1111/1346-8138.15304. PMID: 32216090.
  10. Maddali Bongi S, Del Rosso A, Mikhaylova S, et al. Efficacy of connective tissue massage and Mc Mennell joint manipulation in the rehabilitative treatment of the hands in systemic sclerosis. Clin Rheumatol. 2009;28(10):1167-1173. (rehabilitation reference) PMID: 19568889.