Stroke Rehabilitation:
A Novel Treatment Pays Off

Gary E Cordingley, MD, PhD
Investigators at the University of Alabama at Birmingham made a convincing
case that an unconventional form of stroke rehabilitation is the real thing.
In a landmark study, researchers at the University of Alabama at Birmingham used a randomized controlled trial --
the gold standard method for evaluating the effectiveness of a treatment -- to show that immobilizing the good arm
of stroke patients and intensively exercising the weakened arm actually improved recovery, even when performed
long after the stroke occurred. At one level, randomized controlled trials in the field of rehabilitation medicine have
been so rare that the publication of each and every one should be applauded. At another level, the outcome of
this study is so satisfying in terms of what we think we know about brain physiology (function) that even if the
results turn out not to be true, they ought to be.

A controlled trial is one in which there is a comparison group of patients that is either untreated or is treated
differently. When a controlled trial is also randomized, it means that upon entering the study, participants agree to
be assigned to one group or the other based on the equivalent of a coin-toss. Randomization eliminates bias that
might otherwise come from (knowingly or unknowingly) assigning more promising patients to one group and less
promising patients to the other.

Publishing their results in the March 2006 online issue of Stroke, a medical journal, Edward Taub, PhD, and
co-workers studied 21 patients treated with "constraint-induced movement therapy" (CI) and compared their
outcomes to another 20 stroke patients who received placebo treatment.

In strokes a loss of circulation damages a portion of the brain, resulting in impairment of whatever mental or bodily
function that part of the brain controls. Strokes often cause weakness in an arm with or without concurrent
numbness. Strokes are the leading cause of long-term disability in the U.S.

The researchers included stroke victims in their study who had mild to moderate impairment in use of their
affected arms, but excluded those with severe impairment. The research subjects varied widely in age, averaging
in their fifties. The investigators selected patients whose stroke had occurred a minimum of one year earlier with
an average interval between stroke and treatment of 4.5 years. Patients with concurrent numbness were included,
but those with poor walking or balance were excluded, as were patients with excessive confusion or too much
additional impairment caused by other medical conditions.

The CI treatment was administered over a 2-week span, during which the good arm was immobilized about 90% of
the time with an arm-sling and a hand-splint. CI patients had 10 weekday sessions with therapists, lasting 6 hours
each. During those sessions, patients received one-on-one therapy that was individualized to their needs and
abilities and involved specific, practical tasks of gradually increasing difficulty. The therapists praised patients
each time their performances improved even just slightly. By contrast, placebo-treated patients received a more
general program of physical fitness, cognitive and relaxation exercises over the same schedule.

The abilities of CI and placebo-treated patients were compared in two main ways. In one, the research subjects
were videotaped in the laboratory while attempting specific tasks like holding a book, picking up a glass and
brushing teeth. Their performances were rated by viewers who were purposely not told which treatment the
subject received. The other rating, called the "real world outcome," came from structured interviews of the
patients and their caregivers concerning performance outside the treatment facility.

The researchers found significant improvements in CI-treated patients compared with both their own initial abilities
and those of patients receiving placebo treatment. The CI patients showed a moderate improvement in their
laboratory skills and a large improvement in use of the affected arms in their daily lives. Improvement was still
evident 4 weeks after treatment, and even after 2 years in the 14 of 21 CI patients who could be retested at that

The researchers interpreted the improvement as due to two factors. The first factor, probably more important for
faster gains, was in overcoming "learned non-use" of the weaker arm. The idea is that after a stroke, patients
quickly learn to avoid using the weaker arm to a greater extent than its impairment might warrant, and CI training
forces them to put it back into action. The second suspected factor, developing more slowly, was "neural
plasticity" or actual rewiring of the brain. In neural plasticity surviving brain cells -- previously uninvolved or less
involved in controlling use of the arm -- attempt to make up for the lost brain cells either by creating new contacts
with other brain cells or by modifying the effectiveness of existing links.

In 1992 researchers at the Hammersmith Hospital in London used positron emission tomographic (PET) scans to
examine patterns of brain use in stroke patients. PET scans are good at showing which parts of the brain are
most engaged by specific tasks. Investigators compared PET scans in 10 patients who recovered from a stroke to
those of 10 patients who never had a stroke. In this study subjects repeatedly moved one hand (which in the
stroke patients was the affected hand) while their brains were being scanned. Compared to non-stroke patients,
stroke patients used more areas on both sides of the brain to perform the requested movements, as if the
surviving brain cells were trying to fill in for their fallen comrades.

Taub and collaborators at the National Institute of Neurological Disorders and Stroke used similar methods to
compare patterns of brain activation in 9 CI-treated stroke patients with those in 7 less-intensively treated stroke
patients. In this 2003 study, CI-treated patients showed a shift in the extent to which different parts of the brain
participated in moving the fingers of the weakened hand. Thus, CI treatment seemed to modify the brain pathways
responsible for the finger movements.

(C) 2006 by Gary Cordingley