Authors:

Roberto Casale, MD
Carlo Damiani, MD Amputee
Venessa Rosati, MD

Affiliations:
From the Department of Clinical
Neurophysiology and Pain ORIGINAL RESEARCH ARTICLE
Rehabilitation Unit (RC), Foundation
“Salvatore Maugeri” IRCCS,
Montescano, Italy; San Raffaele
Portuense Tosinvest (CD); and San
Raffaele Pisana (VR), IRCSS, Mirror Therapy in the Rehabilitation
Tosinvest, Roma, Italy.
of Lower-Limb Amputation
Correspondence: Are There Any Contraindications?
All correspondence and requests for
reprints should be addressed to
Roberto Casale, MD, Department of
Clinical Neurophysiology and Pain ABSTRACT
Rehabilitation Unit, “Salvatore
Maugeri” Foundation, IRCCS,
Casale R, Damiani C, Rosati V: Mirror therapy in the rehabilitation of lower-limb
Rehabilitation Institute of amputation: Are there any contraindications? Am J Phys Med Rehabil 2009;88:
Montescano, Via per Montescano, 837– 842.
27040 Montescano (PV), Italy.
Objective: Mirror box therapy and its development (immersive virtual
Disclosures: reality) is used in pain therapy and in rehabilitation of people with ampu-
None. tation affected by phantom limb-related phenomena. It allows patients to
view a reflection of their anatomical limb in the visual space occupied by
0894-9115/09/8810-0837/0
their phantom limb. There are only limited reports of its possible side
American Journal of Physical
Medicine & Rehabilitation effects.
Copyright © 2009 by Lippincott
Williams & Wilkins
Design: We retrospectively reviewed the existence of side effects or
adverse reactions in a group of 33 nonselected patients with phantom
DOI: 10.1097/PHM.0b013e3181b74698 limb-related phenomena.
Results: Nineteen reported confusion and dizziness, 6 reported a not
clearly specified sensation of irritation, and 4 refused to continue the
treatment. Only 4 of the 33 patients did not have any complaints.
Conclusions: Possible reasons for this large number of side effects
could be the lack of selection of patients and the fact that the mirror box
therapy was paralleled by a conventional rehabilitation approach targeted
to the use of a prosthesis. Warnings on the need to select patients, with
regard to their psychologic as well as clinical profile (including time from
amputation and clinical setting), and possible conflicting mechanisms
between mirror box therapy and conventional therapies are presented.
Key Words: Phantom Limb-Related Phenomena, Mirror Therapy, Side Effects

www.ajpmr.com Mirror Therapy 837

 I n recent years, many articles have been published
indicating the growing interest of various research
Mini Mental State Examination normalized for age
and education, and the absence of severe psychiat-
ric disorders such as major depression according to
groups all over the world toward the so-called Diagnostic and Statistical Manual of Mental Disor-
mirror box therapy (MBT) in neurology,1 pain ther- ders IV criteria.
apy,2 and in the rehabilitation of amputated pa- All patients started treatment between 20 and
tients affected by phantom limb-related phenom- 30 days after the amputation had been performed.
ena (PLRP).3– 8 The term PLRP encompasses all the MBT was applied to all participants in 30-minute
positive and negative sensations felt by people with daily sessions (from Monday to Friday) for 3 wks.
amputation in their missing limb, including phan- The MBT technique used was simple and based on
tom limb sensations and phantom limb pain.9 descriptions in the existing literature.2,3,7 The pa-
The mirror box is a device that allows people tient was seated comfortably in front of a modular
with amputation to view a reflection of their ana- box with a mirror placed with its reflecting surface
tomical limb in the visual space occupied by their medially to the remaining limb. When the patient
phantom limb. The box has been reported to in- looked at himself/herself in the mirror, the image
duce vivid sensations of movement originating reflected appeared with both limbs. The patient was
from the missing limb, improving motor control then asked to make a flexion-extension movement
over a paralyzed phantom limb,6 and to reduce (90 –180 degrees) of the remaining limb and at the
PLRP.3–5,7 It has been suggested that MBT may same time to try thinking that he/she was also
work by providing a means to link the visual and moving the amputated limb.
motor systems to recreate a coherent body image All of the patients also followed a parallel tra-
and update internal models of motor control.4 The ditional rehabilitation program that included exer-
basis of this possible effect has been assumed to be as cises for control of the trunk, for improving stump
a result of the activation of mirror neurons in the muscle strength, occupational therapy, and pro-
hemisphere of the brain that is contralateral to the gressive training in the use of a prosthesis.
amputated limb,10 firing when a subject observes an Given that MBT is a relatively new approach,
action or observes someone else performing it.7 A all patients were asked to sign informed consent
development of MBT is the use of virtual reality tech- about the method and purpose of the treatment in
nologies in which the patient does not look into a which it was clearly stated that they were free to
mirror but looks at a personal computer screen in interrupt the treatment at any moment. They all
which the movements of the anatomical limb are agreed to undergo the proposed treatment.
transposed into the movements of a virtual limb and
presented in the space of the person’s missing limb.11 Retrospective Data Assessment
Side effects of this type of therapy have never This retrospective study was approved by the
been defined, and only sparse reports are present in internal review board. Data collection was focused on
literature suggesting that some degree of unease the evaluation of any adverse reactions to the mirror
can be induced by the application of MBT.11 How- therapy that were reported in a clinical diary or ob-
ever, in our personal experience, we have observed served from the recordings of the psychologic inter-
collateral adverse effects to the application of MBT, views. All subjects were followed throughout the
such as lack of appreciation or even displeasure. treatment by a weekly psychologic working group. All
This prompted us to review retrospectively the ex- subjects were free to express openly their feelings
istence and severity of side effects or adverse reac- related to the amputation, to the quality and efficacy
tions induced by MBT in a group of unselected of the mirror therapy, and to the rehabilitation treat-
people with PLRP after amputation. ment in general. Retrospectively reviewed data in-
cluded the number of patients who completed the
MATERIALS AND METHODS MBT, the number of withdrawals, the number of daily
In January 2007, we introduced MBT into the treatments before withdrawal, and the reason for
inpatient rehabilitation process for people with stopping the therapy, with special attention to the
lower-limb amputation with PLRP. So far, we have quality of the reported adverse reactions and their
treated 33 patients between the ages of 18 and 90 intensity, expressed on a 0 –10 scale at the time of the
yrs (9 women; 24 men), amputated at different withdrawal.
levels (transfemoral 20, transtibial 9, and other
levels 3) because of trauma, peripheral arteriopa- RESULTS
thy, cancer, or infection (Table 1). The selection In this unscreened group of people with ampu-
criteria included agreement to participate in a psy- tation, of 33 patients only 4 did not withdraw and
chological working group consisting of free verbal completed the 15 days of prescribed MBT. Ten pa-
communication with a semistructured form, the tients dropped out immediately after the first session,
absence of cognitive impairment, as assessed by the 10 patients after 2 days of treatment, 5 patients after

838 Casale et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 10, October 2009
 TABLE 1 Demographic and clinical data of a nonselected group of 33 lower-limb amputees
affected by phantom limb-related phenomena (PLRP) and adverse reactions induced by
mirror box therapy (MBT)
Demographic Data Clinical Data Adverse Reactions Induced by MBT

Level of Treatment Intensity

Case Age, yrs Sex Amputation Etiology Withdrawal Quality (VAS 0–10)

1 78 M TF PA Yes D 8
2 71 M TF TR Yes D 8
3 73 M TT PA Yes D 10
4 61 F TT PA No NA NA
5 58 M TF PA Yes D 7
6 75 M TF PA Yes D 7
7 57 M TF PA Yes D 7
8 32 M TF TR Yes D 6
9 73 M TF PA Yes I 7
10 70 M TF PA Yes D 7
11 74 F TF PA Yes I 10
12 40 M TT PA Yes D 7
13 26 F OL PA Yes I 10
14 73 M TF IN Yes I 7
15 55 M TF PA Yes D 10
16 84 M TF PA Yes U 6
17 30 M TT PA Yes D 7
18 75 M TF TU Yes D 6
19 53 M TF PA No NA NA
20 71 F TF PA Yes U 6
21 66 M TF PA Yes D 7
22 86 F TF PA No NA NA
23 73 F OL IN Yes U 6
24 74 M TF PA Yes D 6
25 67 M TT PA No NA NA
26 64 F TT PA Yes D 7
27 77 F TF TR Yes I 7
28 71 M TT TR Yes D 6
29 65 M TF TR Yes U 6
30 72 M TF PA Yes D 7
31 75 F TF IN Yes I 7
32 80 M TF TR Yes D 8
33 70 F TF PA Yes D 8
Mean ⫽ 65.7 M ⫽ 23 (69.7%) TF ⫽ 25 (72.7%) IN ⫽ 3 (9.1%) NO ⫽ 4 (12.1%) D ⫽ 19 (65.5%) Mean ⫽ 7.3
SD ⫽ 14.9 F ⫽ 10 (30.3%) TT ⫽ 8 (21.2%) PA ⫽ 23 (69.7%) Yes ⫽ 29 (87.9%) I ⫽ 6 (20.7%) SD ⫽ 1.3
OL ⫽ 2 (6.1%) TR ⫽ 6 (18.2%) U ⫽ 4 (13.8%)
CA ⫽ 1 (3%)

TF, transfemoral; TT, transtibial; OL, other level; IN, infection; PA, peripheral arteriopathy; TR, trauma; CA, cancer; D,
dizziness; I, irritation; U, uneasiness; NA, not applicable; VAS, visual analog scale.

3 days of treatment, and 4 patients after 5 days of somatosensory and motor cortex of adult brain.12–15
consecutive therapy. Table 1 shows the number and These plastic changes have fundamental implications
percentage of patients who withdrew, the quality of for understanding chronic pain and developing new
the adverse reactions, and their intensity. Details of rehabilitation strategies in patients with neurologic
the withdrawals are also shown in Figure 1. No rela- diseases.16 Cortical reorganization in people with am-
tion was seen between withdrawal and sex or age of putation was first postulated and then demonstrated
the patient or level of amputation. by Ramachandran et al.1,17,18 The application of MBT
PLRP consisted mainly of exteroceptive and pro- in controlling PLRP relies on these demonstrations.
prioceptive sensations such as itching, pressure, pain, It has been suggested that the mirror box may work
movement, and abnormal thermal sensations. None in some patients by providing a means to link the
of the patients openly reported an increase or wors- visual and motor systems in helping them recreate a
ening in the PLRP because of the mirror therapy as a coherent body image and update internal models of
cause of their withdrawal. No physical irritation re- motor control.4 To achieve this, the user must at-
lated to the movements was reported. tempt to ignore the intact limb providing the reflec-
tion and focus on the image reflected in the mirror.11
DISCUSSION Despite the great rehabilitation interest, so far
It has been convincingly demonstrated that only a few articles have been published on the
functional reorganization can take place in the clinical results of this treatment, reporting a very

www.ajpmr.com Mirror Therapy 839

 FIGURE 1 Adverse reactions induced by mirror box therapy (MBT) in a nonselected group of 33 lower-limb
amputees affected by phantom limb related phenomena (PLRP).

limited and selected number of treated cases.3,11,18 ferent and perhaps interfering factors. The most
As far as we know, the clinical limitations and side intriguing possibility is that the parallel conven-
effects have never been defined, and only occa- tional rehabilitation approach, mainly targeted to
sional side effects have so far been reported. In one the use of a prosthesis, could have negatively in-
report including only three patients, the major terfered with the mirror therapy, but other concur-
complaints were described as sickness, uneasiness, ring factors could also be hypothesized, such as the
and odd sensations. One patient reported a strong fact that the group of patients treated was un-
sensation of tiredness after the sessions.11 In the screened. The time passed from amputation and
only extensive study present in the literature, in the clinical setting in which the MBT was per-
which 80 people with lower-limb amputation un- formed could also have played some role.
derwent MBT with substantially negative results, Fundamental points in the rehabilitation of
no side effects or withdrawal from therapy was amputees are the elaboration of coping strategies
reported. In this research, a number of patients toward the limb loss, the recognition of the impair-
only described a certain degree of surprise in see- ment, and the acceptance of prosthetic devices.20
ing their missing limb, without clarifying the na- In these processes, the combined action of any
ture, positive or negative, of this surprise.19 sensory inputs can induce modifications in phan-
Although our retrospective analysis has some tom limb awareness and in PLRP, which can be
recognized limitations, such as the fact that the stud- perceived more vividly and sometimes even more
ied cohort was unscreened and not profiled psycho- distressfully.21 Indeed, side effects were reported by
logically, this is the first report on clear difficulties in a patient whose anatomical leg accidentally touched
the extensive application of MBT in people with the prosthesis during a session. This contact pro-
PLRP. The side effects of nonselective application of duced what the subject defined as a “queer feeling.”11
MBT, reported here, strongly challenge the notion Moreover, somatosensory manipulation of painful
that this therapy can always help people with ampu- muscle areas in the contralateral limb also seem to be
tation to recreate a coherent body image and update able to modify PLRP,22 further supporting the impor-
internal models of motor control.4 tance of any sensory inputs and indicating that even
The large number of side effects observed in subtle changes in sensory information from both
our cohort of patients could be as a result of dif- body sides can affect phantom limb awareness. Our

840 Casale et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 10, October 2009
data can extend this concept to a more generalized patients could be even more pronounced in upper-
change in congruency felt by the subjects as confu- limb amputees because of the wider cortical repre-
sion, dizziness, irritation, and unwillingness to con- sentation and, consequently, the more profound
tinue a therapy perceived as disturbing. and extensive reorganization of the somatosensory
The perception of body image in a person with cortex after upper-limb loss.12,13,15
amputation is altered by the convergence of several In this preliminary retrospective study, quali-
factors: the fragmentation or overlapping of several tative findings provide sufficient proof of principle
pieces of information regarding the visual experi- to justify further investigation of MBT and related
ence of the missing limb or both23; the awareness techniques in controlled trials to define better the
of a new, although distorted, sensory reality21; and pros and cons of these treatments. This report
the reconstruction of a body image embracing the highlights the importance of comparing MBT vs.
prosthesis24 induced by the traditional rehabilita- traditional rehabilitation approaches and the cre-
tion process. ation of homogeneous groups of patients, subdi-
A prosthetic limb can be considered as a so- vided according to age, time since amputation, and
phisticated tool, and higher primates25 and hu- types of sensations perceived. This last point is of
mans26 have the capacity to “incorporate” tools pivotal importance because some people with lower-
into their body’s brain representation. Moreover, limb amputation describe their PLRP as specific
the use of a prosthesis or actuator may induce somatosensory alterations whereas others report
cortical and subcortical remapping of such a mag- only a general awareness of the missing limb.29 A
nitude to be detectable using magnetic resonance crucial issue for the development of this technique
imaging studies.27 These experimental data sup- is, therefore, the identification of appropriate psy-
port the clinical improvement in PLRP in patients chophysical clinical evaluation items to ascertain
wearing a functional prosthesis.24,27 what kind of alterations of the body schema or body
During a session of mirror therapy, the patient awareness is more suitable for this treatment.30
has to carry out active exercises and to observe The need for tailored psychologic support and the
them using a mirror, providing the illusion of possible influence of the therapeutic setting are
having two complete, healthy limbs in synchro- implicitly highlighted by a very recent case report,
nized movement (deception of the mirror). This in which 3 mos of home-delivered MBT was suc-
can be interpreted as confounding and in contrast cessfully applied together with continuous psycho-
to the inner process of elaborating the loss and the logic support in a very young and cooperative pa-
parallel incorporation of the prosthesis into the tient.31
body schema. If the mirror method is integrated
parallel to traditional methods, such as simply ACKNOWLEDGMENTS
wearing the prosthesis or sensory stimulation of We thank Dr. Alberto Esquenazi for his en-
the stump,8,20 it could become impossible to guar- couragement and support.
antee a coherent rehabilitation process because the
two methodologies are clearly conflicting: the mir-
ror therapy is trying to trick the brain into think- REFERENCES
ing that the phantom limb is moving normally as it 1. Ramachandran VS: Plasticity and functional recov-
ery in neurology. Clin Med 2005;5:368 –73
was before the amputation, and the traditional re-
habilitation is trying to complete the reconstruc- 2. Moseley GL: Graded motor imagery for pathologic
tion of a new body image that includes a prosthesis. pain: A randomized controlled trial. Neurology 2006;
Time from the amputation should also be con- 67:2129 –34
sidered as a possible interfering factor when apply- 3. MacLachlan M, McDonald D, Waloch J: Mirror treat-
ing MBT. It is recognized that early rehabilitation ment of lower limb phantom pain: A case study. Dis
and prosthesization are positive prognostic fac- Rehabil 2004;26:901– 4
tors.20 However, no data are available in literature 4. Murray CD, Patchick E, Pettifer S, et al: Immersive
on the best time to apply MBT: should it be used virtual reality as a rehabilitative technology for
early or should it be used only in the case of poor phantom limb experience: A protocol. Cyberpsychol
results with more traditional approach? In a very Behav 2006;9:167–70
recent study, the rate of prosthesis acceptance was 5. Murray CD, Patchick EL, Caillette F, et al: Can
16 higher in patients treated early, indirectly sug- immersive virtual reality reduce phantom limb pain?
gesting the possibility of easier neural remodeling Stud Health Technol Inform 2006;119:407–12
in the early postamputation period.28 Nevertheless, 6. Brodie EE, Whyte A, Waller B: Increased motor
the issue of how early MBT should be applied control of a phantom leg in humans results from the
deserves further investigation. Our cohort of pa- visual feedback of a virtual leg. Neurosci Lett 2003;
tients were all lower-limb amputees. It is reason- 341:167–9
able to suppose that what we observed in these 7. Chan BL, Witt R, Charrow AP, et al: Mirror therapy

www.ajpmr.com Mirror Therapy 841

 for phantom limb pain. N Engl J Med 2007;357: upon phantom limb pain, sensation and movement.
2206 –7 Eur J Pain 2007;11:428 –36
8. Giummarra MJ, Gibson SJ, Georgiou-Karistianis N, et 20. Casale R, Alaa L, Mallick M, et al: Phantom limb-
al: Central mechanisms in phantom limb perception: related phenomena and their rehabilitation after
The past, present and future. Brain Res Rev 2007;54: lower limb amputation. Eur J Phys Med Rehabil
219 –32 February 26, Epub ahead of print
9. Mishra S, Bhatnagar S, Gupta D, et al: Incidence and 21. Hunter JP, Katz J, Davis KD: The effect of tactile and
management of phantom limb pain according to visual sensory inputs on phantom limb awareness.
World Health Organization analgesic ladder in am- Brain 2003;126:579 – 89
putees of malignant origin. Am J Hosp Palliat Med 22. Casale R, Ceccherelli F, Labeeb AA, et al: Phantom
2007;24:455– 62 limb pain relief by contralateral myofascial injection
10. Iacoboni M, Mazziotta JC: Mirror neuron system: with local anaesthetic in a placebo-controlled study:
Basic findings and clinical applications. Ann Neurol Preliminary results. J Rehabil Med 2009;41:418 –22
2007;62:213– 8 23. Melzack R: Phantom limbs and the concept of neu-
11. Murray CD, Pettifer S, Howard T, et al: The treat- romatrix. Trends Neurosci 1990;13:88 –92
ment of phantom limb pain using immersive virtual 24. Weiss T, Miltner WH, Adler T, et al: Decrease in phan-
reality: Three case studies. Disabil Rehabil 2007;29: tom limb pain associated with prosthesis-induced in-
1465–9 creased use of an amputation stump in humans. Neu-
12. Wall JT, Xu J, Wang X: Human brain plasticity: An rosci Lett 1999;272:131– 4
emerging view of the multiple substrates and mech- 25. Hihara S, Notoya T, Tanaka M, et al: Extension of
anisms that cause cortical changes and related sen- corticocortical afferents into the anterior bank of the
sory dysfunctions after injuries of sensory inputs intraparietal sulcus by tool-use training in adult
from the body. Brain Res Rev 2002;39:181–215 monkeys. Neuropsychologia 2006;44:2636 – 46
13. Feldman DE, Brecht M: Map plasticity in somatosen- 26. Maravita A, Spence C, Driver J: Multisensory inte-
sory cortex. Science 2005;310:810 –5 gration and the body schema: Close to hand and
14. Killackey HP, Rhoades RW, Bennett-Clarke CA: The within reach. Curr Biol 2003;13:531–9
formation of a cortical somatotopic map. Trends 27. Maruishi M, Tanaka Y, Muranaka H, et al: Brain
Neurosci 1995;18:402–7 activation during manipulation of the myoelectric
15. Chen R, Cohen LG, Hallett M: Nervous system reor- prosthetic hand: A functional magnetic resonance
ganization following injury. Neuroscience 2002;111: imaging study. Neuroimage 2004;21:1604 –11
761–73 28. Biddiss EA, Chau TT: Multivariate prediction of up-
16. Flor H: Cortical reorganisation and chronic pain: per limb prosthesis acceptance or rejection. Disabil
Implications for rehabilitation. J Rehabil Med 2003; Rehabil Assist Technol 2008;3:181–92
Suppl 41:66 –72 29. Mayer A, Kudar K, Bretz K, et al: Body schema and
17. Ramachandran VS, Stewart M, Rogers-Ramachand- body awareness of amputees. Prosthet Orthot Int
ran DC: Perceptual correlates of massive cortical 2008;1:1–20
reorganization. Neuroreport 1992;3:583– 6 30. Gallagher P, Horgan O, Franchignoni F, et al: Body
18. Ramachandran VS, Rogers-Ramachandran D: Syn- image in people with lower-limb amputation: A Ra-
aesthesia in phantom limbs induced with mirrors. sch analysis of the Amputee Body Image Scale. Am J
Proc Biol Sci 1996;263:377– 86 Phys Med Rehabil 2007;86:205–15
19. Brodie EE, Whyte A, Niven CA: Analgesia through 31. Darnall BD: Self-delivered home-based mirror ther-
the looking-glass? A randomized controlled trial in- apy for lower limb phantom pain. Am J Phys Med
vestigating the effect of viewing a “virtual” limb Rehabil 2009;88:78 – 81

842 Casale et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 10, October 2009