Background: An estimated 55 to 75 percent of the 566,000 patients who survive a stroke each year in the United States have ongoing upper extremity limitations at three to six months after stroke. Repetitive training of the affected limb could improve function, particularly by restraining the functional, unaffected side for a period. Constraint-induced movement therapy (CIMT) was the mainstay of the Extremity Constraint Induced Therapy Evaluation (EXCITE) trial. In this randomized, single-blinded trial, Wolf and colleagues tested CIMT on patients who retained some finger and hand movement three to nine months after a stroke.

The Study: Participants were recruited from seven clinical sites after experiencing a first-time ischemic stroke in the previous three to nine months. The 222 eligible participants were randomized to usual care (n = 116) or CIMT (n = 106), with 169 completing all follow-up. Demographic, stroke-related, and cognitive characteristics were similar between groups.

Eligible participants were divided into higher-functioning and lower-functioning groups, depending on the degree of hand function retained. Inclusion criteria included ability to balance and transfer while wearing the restraint and ability to perform several other activities. Participants were randomized to CIMT or usual care, which ranged from no treatment, to various mechanical devices, to occupational or physical therapy.

CIMT patients were instructed to wear a constraining safety mitt 90 percent of their waking time for 14 consecutive days. These patients received behavioral motor training for adaptive and standard task practice based on usual functional activities. Outcome measures included performance on a motor function test (Wolf Motor Function test [WMFT]) and a real-world arm use test (Motor Activity Log [MAL]). Secondary outcome measures included score on the Stroke Impact Scale (SIS) and several other measures. Patients were evaluated five times during the year-long study.

Results: Immediately after treatment, the CIMT group showed improvement of the impaired limb for almost all outcomes compared with the control group. The only exception was performance on two strength items on the WMFT; however, intervention patients had greater improvement on these items than the control group at the 12-month follow-up.

Significant improvements in the intervention group at 12 months were noted in WMFT performance time and MAL scales but not on the WMFT functional ability scale. CIMT participants also had significantly greater gains in SIS hand function than the control group but not in other areas of physical function. The level of baseline hand functioning did not have an effect on treatment.

Intervention patients were able to perform a greater number of WMFT tasks over time compared with control patients, and thus met the authors' definition of clinically relevant improvements. By 12 months, however, clinically relevant improvement attenuated. At 12 months, CIMT participants had a 24 and 65 percent improvement in two use parameters (more use and independent use, with cutoffs defined as better-than-50 percent improvement on the MAL) compared with baseline, indicating a significant improvement compared with control participants.

Conclusion: Intervention and control groups showed improvements in paretic arm function three to six months after stroke, but the CIMT group showed significantly greater improvements. These improvements were greatest immediately after treatment, but gains in a number of outcome measures persisted for 12 months, indicating a specific benefit of CIMT.