Distorted body image in complex regional pain syndrome.

Regional anesthesia results in shrinkage of the primary sensory cortex (S1) representation of the area and the perception that the area is larger than it is.1 Complex regional pain syndrome type 1 (CRPS1) also involves shrinkage of S1 representation2 and, anecdotally, the perception of marked swelling when none is apparent. We posited that if a reduced S1 representation of the affected limb is involved in generating a perception that the limb is larger than it really is, then this effect should be present in patients with CRPS1. Methods. Fifty patients diagnosed with CRPS13 initiated by wrist or hand fracture and 18 patients with non-CRPS1 hand or wrist pain were eligible (see table E-1 on the Neurology Web site at www.neurology.org). Exclusion criteria were pain elsewhere (five patients with CRPS1), symptoms extending beyond the affected limb (four patients), psychiatric diagnosis (two patients), and unable to understand English (one patient). A 3,000-DPI, 4.8 3.2-cm digital photograph was taken of the two hands and distal one-third of the forearms, placed side by side. The image of the affected limb was compressed or expanded in one dimension to 85%, 90%, 95%, 100%, 105%, 110%, or 115% such that an expanded image made the limb look thicker but not longer than it was. Seven 4.2 2.8-cm images of the limb pairs (each incorporating one of the thickness manipulations of the affected hand) were positioned randomly on a 19-inch 1,280 1,024 resolution color monitor. Patients selected the photograph they believed to be accurate. Pilot data showed that image pair selection is reliable in patients with non-CRPS1 pain (intercorrelation coefficient [ICC] 0.93). The size of the affected limb was estimated by the ratio between limbs of the mean circumference taken midway along the proximal phalanx of fingers 2 to 4, using hand measuring tape (Beiersdorf-Jobst, Hamburg, Germany). This measure is reproducible (ICC 0.9). A Mann-Whitney U test was used to test the difference between the perceived sizes of the affected limb selected by subjects and controls. To identify whether the selected image was related to patient characteristics, linear regression between the selected image and finger circumference ratio, mean pain intensity, age of the patient, and duration of symptoms, with correction for multiple measures, was used. Assessors were blinded to the purpose of the study. Logistic regressions were run testing the relationship of the perceived relative size of the affected limb to the presence or absence of each medication, the presence of apparent swelling (to the investigator), and apparent atrophy (to the investigator). Patients gave informed consent. Procedures were approved by the institutional ethics committee and conformed to the Declaration of Helsinki. Results. For patients with CRPS1, the median selection showed that the affected limb expanded to 105% of the actual width. The mean (SD) size of the affected limb in the selected image was 107% (3%). Sixty-three percent of patients with CRPS1 and 17% of the control group selected an image that showed the affected limb expanded (figure E-1). The selected image did not relate to finger circumference, pain intensity, or age (p 0.3 for all) but did relate to duration of symptoms (r 0.55, F(1,36) 15.5, p 0.001). There was no effect of medication, apparent swelling, atrophy, or side of affected limb between patients who selected enlarged images and those who did not (p 0.41). In control patients, the median selection showed the affected limb at 100% and the mean (SD) width of the limb in the selected image was 100% (2%), which was smaller than that for patients with CRPS1 (Mann-Whitney U 136, Z 3.7, p 0.001). Discussion. Patients with CRPS1 perceived the affected limb to be larger than it really was. This distortion of body image may be an important part of the presentation of CRPS1. The mechanisms involved are not clear, but S1 changes may be involved. S1 changes have been proposed to underpin clinical phenomena (e.g., stimulation at one site referring sensation to body parts represented immediately adjacent to the affected area in S14) that occur in patients following amputation, tooth extraction, spinal cord injury, stroke, local anesthesia, and CRPS1. Those groups also describe distortion of body image such that the anesthetized part feels large, full, or swollen. Perhaps through neural connections indirectly linking S1 and association cortices in the “what” visual pathway, shrinkage of the S1 representation of the affected limb engenders alteration of visual magnitude perception. It is notable that the perceived size of the affected limb related to the duration of CRPS1, and in other patient groups characterized by changes in S1 representation, e.g., amputees, cortical reorganization correlates with the duration of symptoms.5 The current sample was not homogeneous; symptoms, anatomic site, and medications varied. Although these factors did not relate to selected image, future work could elucidate the importance of such factors and verify the current results. The results corroborate other findings of distorted body image in people with CRPS1, but the mechanisms remain poorly understood.