In 1871, Mitchell1 used
the term "phantom limb" to describe the commonly reported
sensation that an amputated limb is still present. In the intervening
years, many studies have been performed to evaluate pain, including
phantom-limb pain, after amputation2-8.
Those studies excluded patients who had the procedure because of
debilitating pain, and therefore they cannot be used to assess the
effectiveness of amputations performed for pain relief. Today, surgeons
are reluctant to consider amputation of a chronically painful extremity
because of the belief that pain will persist or intensify after
amputation. Katz and Melzack4 based
their concept of "somatosensory memories" on a
series of eight patients who had undergone amputation for undescribed
indications, six of whom described phantom-limb sensations that
were similar to preamputation pain. Other reports have seemingly
corroborated the findings of Katz and Melzack; however, those studies
generally excluded patients who requested elective amputation for
the treatment of intractable pain2,3,5-15.
The purpose of the present study was to assess the outcome of below-the-knee
amputations performed in selected patients who had intractable foot
and ankle pain. Our hypothesis was that patients would have significant
relief of pain and a measurable increase of functional capability
after amputation.
The study was approved by the institutional review board at the
institution of the lead investigator (C.L.S.) and was performed
at three participating institutions. Patients were included in the
study if (1) a below-the-knee amputation had been performed because
of intractable foot or ankle pain, (2) maximal medical therapy had
failed to alleviate the pain, (3) the limb was considered surgically
unreconstructible by an orthopaedic foot and ankle specialist, and
(4) there was a minimum follow-up period of twenty-four months after
the amputation. Patients with diabetes mellitus, peripheral vascular occlusive
disease, or peripheral neuropathy were excluded. Each patient had
pain that was limited to the foot and ankle area, which allowed
a standard below-the-knee amputation. There were no surgical complications,
and a standard posterior flap was used in all patients.
All patients were contacted initially so that they could be informed
of the study and asked to participate. After verbal consent had
been received, a questionnaire regarding surgical outcome was either
sent to the patient or administered over the telephone.
Questionnaire
The questionnaire consisted of two similar parts, each of which
was two pages long. The first part focused on preoperative pain
and function, and the second part focused on postoperative pain
and function. The questionnaire was adapted from standardized instruments16-18 and was used to assess the patient’s
level of pain, need for analgesic medication, days missed from work,
recreational status, satisfaction with the procedure, and willingness
to have the amputation performed again under similar circumstances. The
questionnaire required short written (or verbal) answers; "yes," "no," or "not
applicable" answers; and use of standardized 10-cm visual
analog scales that had positive and negative word anchors on opposite
ends. A score of 0.0 represented a favorable outcome ("no
pain" or "completely able"), whereas a
score of 10.0 represented a poor outcome ("constant pain" or "completely
unable").
The two-part questionnaire consisted of the following sections.
General
The questionnaire assessed the patient’s general response
to the below-the-knee amputation, perceived disability (with a score
of 0.0 representing "completely normal" and a
score of 10.0 representing "completely disabled"),
and work status. The patient was specifically asked whether he or
she was "satisfied with the results" of the amputation;
whether he or she would, "under similar circumstances,
have the amputation done again"; and whether any other
treatment had been required since the time of the amputation.
Pain
The first part of the questionnaire asked the patient to rate, with
use of a visual analog scale, the frequency and intensity of preoperative
pain as well as the degree to which the pain had hindered daily
life before the amputation (with 0.0 representing "never" and
10.0 representing "completely"). The second part
of the questionnaire asked the patient to rate, with use of a visual
analog scale, the frequency and intensity of phantom-limb pain,
the frequency of phantom-limb sensations, and the degree to which
the phantom-limb pain hindered daily life at the time of the latest
follow-up.
The patients’ medical records were reviewed for the
six-month period immediately prior to the amputation and for the six-month
period immediately prior to the last follow-up. From these
records and from the responses to the completed questionnaires,
we determined, for each patient, the number of physician visits
and the usage of analgesic medications for leg or foot pain. We
stratified medication use according to the United States Drug Enforcement
Agency’s schedule for narcotic classification.
Recreation
Both the preoperative and postoperative portions of the questionnaire
asked the patient to rate his or her ability to use public facilities
(restaurants, malls, and so on) and to estimate how often he or
she was forced to withdraw from social activities because of severe
pain. Both portions also included questions on walking distance,
the ability to drive a car, and the level of ability to participate
in sports and recreational activities.
Statistical Methods
The Wilcoxon rank-sum test was used to determine whether there
were significant differences between the preoperative and postoperative
scores for pain, recreation, and general disability. The McNemar
test was used to analyze any difference between the preoperative
and postoperative responses regarding the use of pain medication,
the ability to drive a car, the ability to participate in sports
and recreational activities, and employment status. The Spearman
coefficient was used to determine any significant correlations between
the variables of age, gender, and duration of follow-up.
A p value of <0.05 was considered significant.
Of the twenty patients who met the inclusion criteria, eighteen were
located and agreed to participate. Fourteen patients completed the
written questionnaire, and four were evaluated with use of a telephone
interview. The four women and fourteen men had an average age of
forty-two years (range, twenty-six to sixty-one
years) at the time of the latest follow-up, which was performed
at an average of forty-one months (range, twenty-five
to eighty-five months) after the amputation. The patients
had had an average of 6.2 operations (range, one to twenty operations)
before the amputation. The history of each patient’s injury
and diagnosis before amputation, as recorded from the hospital charts,
is listed in the Appendix.
Two patients were not included in the study. The first patient was
a man who had the amputation because of recurrent painful infection
and ulceration related to acromegaly and overgrowth of osseous elements.
He was followed for approximately three years, during which time
the pain decreased and function greatly improved. The patient then moved
out of state and was lost to follow-up. Attempts to locate
this patient were successful, but he declined to be interviewed.
The second patient was a woman who was doing well at the time of
her initial enrollment in the study, eleven months after the amputation.
She was initially able to run while wearing the prosthesis but then
became dissatisfied with her performance and sought additional care.
After the patient found several custom-made prostheses to be unacceptable, she
had an Ertl procedure with resection and proximal transposition
of a tibial neuroma stump approximately thirty months after the
amputation. She subsequently moved out of state but, at the time
of writing, reported that she was pleased with her outcome and was
able to perform her activities without pain.
General
When the patients were asked whether they would have the below-the-knee
amputation again under similar circumstances, sixteen said yes,
one was unsure, and one said no. The same distribution was observed
when the patients were asked whether they were satisfied with the
outcome (sixteen said yes, one was unsure, and one said no).
Patients reported a decrease in their overall disability, with
the average score improving from 8.4 to 3.7 (p = 0.0001).
Three patients had worked full time before the amputation, and all three
returned to full-time work after the amputation. In addition,
five patients who had been unable to work before the amputation
gained full-time employment postoperatively. The main reasons
mentioned by the ten patients who did not return to full-time
employment included other health problems (four patients), increased
but continued pain and disability related to the amputated limb
(two patients), prosthetic complications (two patients), an inability
to walk long distances (one patient), and a chronic shoulder injury
(one patient).
Two patients required additional local care after the below-the-knee
amputation. One patient had resection of the proximal part of the
fibula because of painful proximal tibiofibular syndesmotic instability
and related difficulties with prosthetic fitting. A second patient
required dermatological treatment of a rash that developed on the
stump.
Patients reported that the most difficult problem that they faced
after the amputation was maintaining an adequately fitted prosthesis.
Pain
Visual-analog-scale scores were used to compare the patients’ preoperative
and postoperative perceptions of pain. Patients reported a decrease
in the degree to which their pain hindered their activities, with
the average score improving from 8.8 preoperatively to 1.3 postoperatively
(p < 0.0001). They also reported a decrease in both pain
frequency (with the average score improving from 9.8 to 1.7; p < 0.0001)
and pain intensity (with the average score improving from 8.4 to
2.6; p < 0.0001). We calculated an "overall pain" measure
for each patient as the product of the frequency and intensity scores. Patients
reported a decrease in overall pain, with the average score improving
from 82 (of a possible 100) preoperatively to 8.7 postoperatively
(p < 0.0001).
The patients required an average of 6.6 office visits for pain
or limb care in the six months before the amputation, compared with
an average of 1.9 office visits during the most recent six-month
period (p < 0.0001). Of the nine patients who had been taking
schedule-II narcotics preoperatively, one continued their use postoperatively
at a lower frequency, two required lower doses of a schedule-III
narcotic, and six required no medication. Similarly, of the eight
patients who had been taking schedule-III narcotics preoperatively,
one continued their postoperative use at a lower frequency, two
switched to schedule-IV narcotics, one switched to aspirin, and
four required no pain medication. The one patient who had been taking
ibuprofen preoperatively required no medication at the time of follow-up.
At the time of follow-up, fifteen of the eighteen patients
experienced phantom-limb sensations, with an average frequency rating
of 3.1 on the 10-cm visual analog scale.
Activity
Patients reported an improved ability to walk and to function in
public places after the amputation, with the average score decreasing
from 6.4 to 2.0 (p < 0.0001). They also reported a decreased
tendency to withdraw from social activities because of pain, with
the average score improving from 6.2 to 1.4 (p = 0.0002).
Three patients who had participated in sports and recreational activities
before the amputation continued to do so postoperatively, and six
additional patients resumed sports and recreational activities that
they had discontinued before amputation. The average distance that
the patients were able to walk increased from 0.3 mile preoperatively
to 0.8 mile postoperatively (p = 0.0034). Seventeen patients
had been able to drive before the amputation, and all eighteen were
able to do so after the amputation.
The present retrospective cohort study demonstrated substantial
improvement in patient outcomes after below-the-knee amputations
performed for intractable foot and ankle pain in limbs that were
considered unreconstructible. The patients were generally selected
for treatment with amputation after the surgeon had (1) identified,
in most cases, an anatomically discrete cause of pain, (2) exhausted
all nonoperative means of treatment, and (3) carefully assessed
the patient’s mental and emotional status to determine
the likelihood of improvement after amputation. The latter evaluation
was not done in a standardized manner, but rather as part of the
routine clinical assessment. During this assessment period, there
were other patients with intractable foot and ankle pain who either
were not willing to have an amputation or were considered unlikely to
improve with amputation.
A weakness of the retrospective cohort study design was that we
asked the patients to estimate their preamputation condition at
an average of three years after the procedure. The average preamputation
scores on the visual analog scales regarding the intensity and frequency
of pain were 8.4 and 9.8, respectively, for a computed overall pain
score of 82 (of a possible 100) points. Although these numbers are
subject to recall bias, they cannot be entirely discounted, given
that the major reason for amputation was severe pain. Similar visual analog
scales or ordinal scales have been used in previous studies to compare
preamputation and postamputation pain in a retrospective manner8,13-15,19. The visual analog scale
that we used in this study is reliable and well validated for the
assessment of foot and ankle pain16-18.
Many studies have been performed to assess the levels of pain and
function in amputees4-10,12-15,20,
but none of those studies focused specifically on the results of
below-the-knee amputations performed because of intractable pain
due to a posttraumatic, degenerative, or anatomically focal pathological
condition of the foot or ankle. The relevance of the information
derived from studies of patients with different disease processes,
different indications for amputation, and different levels of amputation
is uncertain.
Similar levels of pain frequency and intensity have been documented
in studies of amputations due to trauma. In an analysis of 2694
amputees who were veterans of the United States military, Sherman
et al.14 reported that 78% of
the individuals had some phantom pain and, of those, 56% reported
that their pain decreased or resolved completely. No correlation
was noted between the preamputation level of pain, the development
of phantom pain, and the ultimate level of pain. In a similar analysis
of 526 amputees who were veterans of the British military, Wartan
et al.8 reported that 55% of
the individuals had some phantom pain and, of those, 53% reported
that the pain decreased or resolved completely. Both of those studies,
however, included both upper and lower extremity amputations as
well as amputations proximal and distal to the knee. Thus, comparability with
the cohort in the present study is limited.
Two previous studies focused on severe pain similar to that reported
by the patients in the present study. Browder and Gallagher20 found that twelve (46%)
of twenty-six patients had persistent phantom-limb pain,
but they provided little qualitative information regarding either
preamputation or phantom-limb pain. Parkes5,
in a study of forty-six patients with both upper and lower extremity
amputations, found that 50% of the patients had persistent
phantom-limb pain. The persistence of phantom-limb pain was correlated
with unemployment, persistent illness, and a rigid, self-reliant
personality.
Despite differences in injury mechanism, amputation level, and
both preamputation and postamputation data, many studies have supported
the notion that the frequency and/or intensity of pain
can decrease after amputation4-6,8,12-15,20.
No studies, to our knowledge, have focused purely on below-the-knee
amputations performed because of intractable foot and ankle pain
in unreconstructible limbs. In addition, none have demonstrated
the dramatic decrease in pain frequency and intensity or the significant
increase in function reported by the patients in the present study.
The differences in our results may be related to patient selection,
the disease process, or consistent skewing of the results within
the relatively small sample size of only eighteen patients. Other
factors, such as personality and social status, have been shown
to influence outcome after amputation12.
It is noteworthy that the patient who reported the worst outcome
in the present study (and answered "no" to the
questions regarding satisfaction with the outcome and willingness
to undergo the procedure again) was the only patient who was involved
in litigation related to the foot or ankle.
The patients in this study did remarkably well. On the average,
they had far less pain, required less medication, and were able
to resume a more normal life after the amputation. Given the chronicity
of the symptoms and the failure of all other surgical and medical
treatments, the results of these amputations are truly remarkable.
Because of the irreversible nature of below-the-knee amputation
and the substantial lifetime impairment with which it is associated,
rigorous prospective studies are needed in order to better predict
improvement in the quality of life of those who might benefit from
the procedure.
A table showing specific demographic and study data is available
with the electronic versions of this article, on our web site at
www.jbjs.org (go to the article citation and click on "Supplementary
Material") and on our CD-ROM (call 781-449-9780, ext. 140,
to order).