Secondary Logo

Journal Logo

Phantom Limb, Residual Limb, and Back Pain After Lower Extremity Amputations

Smith, Douglas G., MD*,**†; Ehde, Dawn M., PhD**‡; Legro, Marcia W., PhD§; Reiber, Gayle E., PhD∥¶; del Aguila, Michael, PhD∥¶; Boone, David A., CP, MPH*†‡#

Section Editor(s): Gottschalk, Frank MD

Clinical Orthopaedics and Related Research®: April 1999 - Volume 361 - Issue - p 29-38
Section I: Symposium: Amputation and New Prosthetic Devices

This study describes the sensations and pain reported by persons with unilateral lower extremity amputations. Participants (n = 92) were recruited from two hospitals to complete the Prosthesis Evaluation Questionnaire which included questions about amputation related sensations and pain. Using a visual analog scale, participants reported the frequency, intensity, and bothersomeness of phantom limb, residual limb, and back pain and nonpainful phantom limb sensations. A survey of medication use for each category of sensations also was included. Statistical analyses revealed that nonpainful phantom limb sensations were common and more frequent than phantom limb pain. Residual limb pain and back pain were also common after amputation. Back pain surprisingly was rated as more bothersome than phantom limb pain or residual limb pain. Back pain was significantly more common in persons with above knee amputations. These results support the importance of looking at pain as a multidimensional rather than a unidimensional construct. They also suggest that back pain after lower extremity amputation may be an overlooked but very important pain problem warranting additional clinical attention and study.

From the Departments of *Orthopaedic Surgery, ∥Health Services and Epidemiology, and ‡Rehabilitation Medicine, University of Washington, Seattle, WA; ¶Veterans Administration Puget Sound Health Care System, Seattle, WA; **Harborview Medical Center, Seattle, WA; †Prosthetics Research Study, Seattle, WA; §MEDTAP International Inc, Seattle, WA; and the #Veterans Administration Center for Amputation, Prosthetics and Limb Loss Prevention, Seattle, WA.

Supported by grants from Department of Veterans Affairs, Health Services Research and Development, and Rehabilitation Research and Development; and the Prosthetics Research Study.

Reprint requests to Douglas G. Smith, MD, Department of Orthopaedics, Harborview Medical Center, 325 Ninth Avenue, Seattle, WA 98104-2499.

Pain after limb amputation is a common sequelae that often becomes chronic, limiting quality of life and functional capacity.15,17 The literature on pain after amputation primarily has focused on phantom limb pain and residual limb pain (also referred to as stump pain). Phantom limb pain has been defined as painful sensations perceived in the missing body part,2 whereas residual limb pain is pain perceived as emanating from the residual portion of the limb (the stump).2 Much less attention has been directed to understanding nonpainful phantom sensations, back pain, and the distinct qualities of intensity, frequency, and bothersomeness that constitute the dimensions of pain.

Phantom limb pain has been reported to be more common than residual limb pain,2,8,13 with reported rates of prevalence ranging from 0.5% to 100%.2,9,17,18 Differences in prevalence rates seem related largely to systematic variations in study methodology, with many of the studies basing their findings on the proportion of patients seen in a medical practice for pain.18 Despite these rate variances, it commonly is thought that most persons who undergo limb amputations, possibly as many as 85%, will experience phantom limb pain.9,17

Residual limb pain is ubiquitous immediately after amputation but generally is thought to subside significantly in a few weeks coincidental with surgical healing.13 Unfortunately, the literature is unclear regarding the prevalence of residual limb pain once the amputation has healed. As reported by Loeser13 and Jensen and Rasmussen,10 much of this ambiguity stems from investigators not carefully distinguishing residual limb pain from phantom limb pain. Contrary to this view, studies that did differentiate between the two pain types found the percentage of individuals with residual limb pain was 22% at 6 months and 21% at 2 years after amputation.8,9

Phantom limb sensations involve the perception of nonpainful sensations in the portion of the amputated limb that is missing. Similar to phantom limb and residual limb pain, research on the prevalence of nonpainful phantom sensations is fraught with methodologic variability, including the fact that nonpainful phantom sensations are not always well differentiated from phantom limb pain.2,8,10 Some estimates suggest phantom limb sensations are experienced by virtually all persons with amputations,13 whereas others report a lower incidence. Jensen et al8 reported that 90% of their prospective sample endorsed phantom limb sensations six months following amputation.

One area of pain after amputation that has not been explored is back pain. Back pain in the general population is common. It has been estimated that between 12% and 45% of the adult population is experiencing low back pain at any one time.1 It is known whether persons with amputations experience back pain similarly to the general population.

Finally, an important methodologic consideration when studying the phenomenon of pain is its definition. In the orthopaedic medical setting, pain traditionally has been thought of as a unidimensional construct, with patients frequently being asked to rate their pain on a 0- to 10-point scale ranging from no pain to severe pain. However, pain researchers view pain as a multidimensional construct with several relevant dimensions, including pain location, intensity, frequency, and affect.6 Pain location involves the bodily location of pain. Pain locations typically measured in studies of amputation related pain are the phantom limb and residual limb. The frequency of pain also commonly is assessed and involves when and how often (duration) pain is experienced. To a lesser extent, pain intensity, which involves how much a person hurts or the severity of the pain, also has been assessed in some studies of postamputation pain. However, there have been significant variations in what types of pain intensity is assessed (worst, average, least). The affective component of the pain experience has been neglected in the amputation pain literature. Pain affect is defined as the emotional arousal and disruption engendered by the pain experience (the bothersomeness). Affect is distinct from intensity; for example, one may have intense pain but not be upset or bothered by it.6

The goal of this paper was to expand on existing knowledge regarding the experiences of pain and sensations after lower extremity amputations by describing the frequency, intensity, and affect (bothersomeness) of nonpainful phantom sensations, phantom pain, residual limb pain, and back pain. Of specific interest were: (1) whether nonpainful phantom sensation occurs with significantly greater frequency and intensity than phantom limb pain; (2) whether phantom limb sensations, phantom limb pain, and residual limb pain decrease with time after amputation; and (3) whether the phantom limb pain is significantly more frequent than residual limb pain or back pain.

Back to Top | Article Outline



Patients who had undergone a lower limb amputation and used a prosthesis were identified from databases in two Seattle area hospitals (an adult trauma center and a Veterans hospital). Of the 293 persons initially identified, 144 patients were located and screened. Inclusion criteria specified that respondents (1) were to be 1 or more years post-unilateral amputation at the Symes level (ankle disarticulation) or higher; (2) use a fitted prosthesis at least 5 days a week; (3) read English; and (4) provide informed consent. Of the 126 individuals who were eligible, 92 consented to participate in the study and returned the questionnaire. The study protocol was approved by the University of Washington Human Subjects Committee.

Back to Top | Article Outline


The Prosthesis Evaluation Questionnaire12 is a 43-item measure developed to measure the prosthesis function and quality of life of persons with lower extremity amputations who use a prosthesis. Prosthesis Evaluation Questionnaire items are grouped into 10 scales that have been shown to have high internal consistency, temporal stability, and content and criterion validity.12 Using the visual analog scale format,4 the Prosthesis Evaluation Questionnaire also has 12 items that assess the frequency, intensity, and bothersomeness of nonpainful phantom sensations, phantom limb pain, residual limb pain, and back pain. The visual analog scale is a 100-mm line bounded by two anchor phrases denoting the extremes of possible answers (never and all the time). Respondents indicated their answer by making a mark across the line. The visual analog scale has been described as one of the most reliable of all existing pain measures19 and is particularly useful in measuring pain intensity, and pain affect.6,15 The Prosthesis Evaluation Questionnaire also asked participants to report their use of nonprescription and prescription medications and alcohol and street drugs for pain.

Types of pain and sensation were defined in the following ways. Phantom limb was defined as the portion of the limb that was amputated or missing. Nonpainful phantom limb sensations were defined as sensations in the missing (phantom) limb that were not painful (the feeling that the missing foot is wrapped in cotton). Phantom limb pain referred to painful sensations in the phantom limb, whereas residual limb pain referred to pain in the portion of the amputated limb that was still physically present (the stump). Back pain was defined as painful sensations in the back, which could include low, mid, and upper back pain.

Back to Top | Article Outline


Potential participants were notified by mail about the project and then received a telephone call to confirm eligibility. Each recruit received a packet by mail containing the Prosthesis Evaluation Questionnaire, a demographic and medical questionnaire, and the consent form. To facilitate subject recruitment, followup calls and mailings were made to those who had not returned the packet.3

The Prosthesis Evaluation Questionnaire visual analogue items were scored using a digitizing tablet with a marking pen that directly entered information into a data management program. This high speed technology was accurate and reduced date entry time over other methods of date entry used for this response format. The distribution of raw scores showed a clustering of responses in the 0 to 4 mm range, and thus for this analysis 0 was expanded to include up to 4 mm.

Back to Top | Article Outline

Statistical Analysis

Descriptive analyses of the visual analog scale items revealed that the Prosthesis Evaluation Questionnaire pain frequency, intensity, and bothersomeness items were not normally distributed. Thus, nonparametric tests for related variables (Wilcoxon signed ranks test, chi square) were used to compare the distributions for the data. When appropriate, correlational analyses were used to explore associations between variables.

Back to Top | Article Outline


Description of Sample

Descriptive analyses were conducted on the 92 participants' responses to the demographic and clinical variables, the results of which can be seen in Table 1. The majority of participants were male (86%) and Caucasian (84%). They ranged in age from 22 to 81 years (mean, 4.9; standard deviation, 13.7). Many (63%) were not employed for reasons that included being retired (43%) and disabled (3.7%). Nearly 51% reported an education of twelfth grade or more.



The most common level of amputation was transtibial (below knee; 63%), with transfemoral (above knee; 25%) being the next most common level of amputation. Trauma accounted for 66.6% of the amputations, with chronic diseases, including diabetes, ulcers, infections, and vascular disease, accounting for 40.7% of the amputations in this sample. (Several subjects attributed their amputation to multiple causes.) The range of time since amputation was 1 to 53 years (mean, 18; standard deviation, 17.2). Although 38% of subjects had undergone amputation within the last five years, an equal number had their amputation more than 20 years ago.

Back to Top | Article Outline

Frequency of Nonpainful and Painful Sensations

Participants were asked to rate how often they experienced each of the four sensations in the four weeks preceding completion of the Prosthesis Evaluation Questionnaire, with the visual analog scale anchors ranging from never to all the time. Analyses of the frequency data, counting those subjects who marked the score five or greater, showed that nonpainful phantom sensations were the most commonly experienced sensation, with 80.4% of the sample reporting nonpainful phantom sensations in the past 4 weeks. Additionally, 63.3% experienced phantom limb pain, 76.1% experienced residual limb pain, and 70.8% experienced back pain in the past 4 weeks. As predicted, nonpainful phantom limb sensations were significantly more frequent than painful phantom limb sensations (Z = -4.84, p < .0001). However, the frequency of phantom limb pain was not significantly different than residual limb or back pain (residual limb: Z = -1.46, p < 0.14; back: Z = -1.53, p < 0.13).

To examine further the frequency of nonpainful and painful sensations, frequency analyses were conducted looking at how often respondents experienced sensations and pain half of the time or more (50-100 mm on frequency visual analog scale). These revealed that 53.3% experienced nonpainful phantom sensations, 36.7% experienced phantom pain, 40.2% experienced residual limb pain, and 32.6% experienced back pain more than half of the time in the preceding 4 weeks.

Surprisingly, time since amputation was not correlated with the occurrence of nonpainful phantom sensations (r = -.02), phantom pain (r = 0.02), residual limb pain (r = -0.11), or back pain (r = 0.12). In fact, of those participants who were more than 3 years postamputation, many reported experiencing nonpainful phantom sensations (75.7%), phantom limb pain (60.9%), residual limb pain (74.3%), and back pain (71.6%) in the preceding month.

To understand better the overall experience of pain after lower extremity amputation, analyses also were conducted looking at the total number of pain problems participants experienced in the 4 weeks preceding completion of the Prosthesis Evaluation Questionnaire. These revealed that only eight participants (9.2%) were pain free during the four-week period assessed. Nearly half (47.1%) experienced all three types of pain (phantom, residual limb, back), whereas 16.1% and 27.6% experienced one and two types of pain problems, respectively, in the preceding 4 weeks.

Back to Top | Article Outline

Intensity of Painful and Nonpainful Sensations

Using the visual analog scale with anchors of extremely mild to extremely intense, participants rated the intensity of the strongest episode of pain or sensation in the preceding four weeks. Their scores then were collapsed into mild, moderate, and severe categories of intensity by dividing the visual analog scale into thirds (0-33=mild). As shown in Figure 1, the intensity of participants' worst episode did not differ significantly among the four domains measured (χ2 = 0.88, p ≤ 0.83). Contrary to hypothesis 1, the intensity of phantom sensations was not significantly different than the intensity of phantom limb pain (Z = -1.07, p ≤ 0.29). Intensity of pain was also not significantly correlated with time since the amputation (r values range from -0.09 to 0.09).

Fig. 1

Fig. 1

Back to Top | Article Outline

Bothersomeness of Painful and Nonpainful Sensations

Using the visual analog scale, with anchors of not at all to extremely bothersome, participants also rated, in general, the bothersomeness of the four types of sensations measured. These scores also were collapsed into mild, moderate, and severe categories of bothersomeness by dividing the visual analog scale scale into thirds (0-33 = mild). As shown in Figure 2, nonpainful phantom sensations were the least bothersome for the sample, with back pain being the most bothersome. As expected, nonpainful phantom sensations were significantly less bothersome than phantom pain (Z = -4.40, p ≤ 0.001), residual limb pain (Z = -5.76, p ≤ 0.001), and back pain (Z = -5.21, p ≤ 0.001). Back pain was significantly more bothersome than phantom limb (Z = -1.91, p ≤ 0.05) but not residual limb pain (Z = -0.25, p ≤ 0.80). Time since amputation was not correlated with the bothersomeness of pain or sensations (r values range from -.04 to 0.21).

Fig 2

Fig 2

Back to Top | Article Outline

Back Pain, Amputation Level, and Age

To explore further the nature of back pain in persons with lower extremity amputations, additional nonparametric analyses were conducted comparing persons with below the knee (n = 54) and above the knee (n = 23) levels of amputation on the frequency, intensity, and bothersomeness of back pain. These showed that persons with above knee amputations were significantly more likely to have had back pain in the previous 4 weeks (Z = -1.99, p ≤ 0.05), to have had greater intensity of pain (Z = -1.99, p ≤ 0.05), and to report their back pain to be more bothersome (Z = -2.15, p ≤ 0.05) than persons with below knee amputations.

Because increasing age is typically associated with the occurrence of back pain in the general population,1 correlational analyses explored the association between age and back pain in this sample. Surprisingly, these analyses showed that age was not significantly correlated with the frequency (r = 0.11), intensity (r = 0.01), or bothersomeness (r = -0.11) of back pain in this sample.

Back to Top | Article Outline

Medication Use

Of the participants (n = 81) reporting at least one of the three types of pain, the majority (63.4%) reported not using any prescription medications for pain in the previous four weeks. Of the 29 participants who did report taking prescription pain medication in the previous 4 weeks, 11 (35.5%) reported taking it on a near daily basis (six or more times/week). Table 2 shows participants' reported use of medications, alcohol, and illicit drugs for the 4 preceding weeks by the three types of pain measured in this study. Contrary to expectations, phantom limb pain did not account for significantly more prescription medication use than did residual limb (Z = -1.27, p ≤ 0.21) or back pain (Z = -1.2, p ≤ 0.23).



As shown in Table 2, participants tended to use over the counter medications more often than they used prescription pain medications. Nonsteroidal inflammatory medications were the most commonly used over the counter medication for phantom (46.2%), residual limb (42.1%), and back (35.3%) pain. Acetaminophens also were used commonly for back pain (29.4%). For phantom (33.3%) and residual limb (50%) pain, the most commonly reported prescription medications used in the past month were the opioid and acetaminophen combination medications (acetaminophen with codeine, oxycodone). Opioid and acetaminophen combinations (30%) and muscle relaxants (carisoprodol, methocarbamol; 40%) were the most commonly reported type of prescription medication used by persons with back pain.

Back to Top | Article Outline


The results of this study suggest that after lower extremity amputations, persons commonly experience various nonpainful and painful sensations and that pain is not limited to the phantom limb. Similar to another report in the literature,8 nonpainful phantom sensations were experienced by many persons with amputations (80% of this sample) and were more frequent than painful phantom limb sensations. A substantial number of respondents also reported experiencing phantom limb (63.3%) and residual limb (76.1%) pain in the preceding 4 weeks. The rate of residual limb pain reported in this sample was notably higher than the rate reported in previous studies8,9 and suggests that residual limb pain may be as common as phantom limb pain.

Although preliminary, these results suggest that back pain may be an overlooked problem after amputation. Back pain was surprisingly prevalent in this sample (71%) and much higher than estimates of back pain in the general population.1 Persons with above knee amputations rated their back pain as more frequent, intense, and bothersome than persons with below knee amputations. Thus, back pain may be related to characteristics of the amputation itself such as the level of amputation or biomechanical variables. Back pain also was rated as significantly more bothersome than the other types of sensations measured. This is surprising in that it suggests that the most emotion provoking and disruptive pain after amputation is not in the phantom limb but in the back.

These results support the notion that pain after amputation is complex and multidimensional in nature. Persons with lower extremity amputations manage pain in not one but several locations. Nearly half of this sample reported experiencing all three types of pain (phantom, residual limb, back) in the preceding 4 weeks; few were pain free. Participants were able to distinguish between the frequency, intensity, and bothersomeness of the four sensations measured and rate them differentially. For example, although phantom sensations were more frequent than phantom pain, they were not significantly more intense. To the authors' knowledge, this is the first study to examine the affective component of pain following amputation.

In this sample, nonpainful and painful sensations persisted years after the amputation. Although nonpainful and painful sensations after amputations typically are thought to decrease with time,10 others have reported a similar persistence in amputation related sensation.11,17 Contrary to expectations, the intensity and bothersomeness of the sensations did not decrease with time.

Although the treatments of phantom limb and residual limb pain are innumerable, there is a lack of good efficacy data for the various pain treatment modalities.5,13 Despite this lack of evidenced based treatment, a sizable minority the patients reported using over the counter and prescription medications for pain. Participants reported using over the counter pain medications more often than prescription pain medications. Phantom limb pain did not account for more prescription pain medication use.

Some cautions are warranted in interpreting these results. The visual analog scale was used to measure the frequency of the sensations in this sample, not allowing for the absolute measurement of pain frequency or pattern. The frequency of pain would be better assessed by asking about the absolute frequency of pain in real time (for example, In the last week, how many days did you experience pain?). Pain intensity was assessed by asking participants to rate their worst episode of pain, which represents only one type of pain intensity. It also would have been useful to measure the least, current, and average intensity ratings.7 The sample only included persons who use a prosthesis and thus may not be representative of all persons with amputations. However, this sample may be more reflective of persons with amputations than some of the studies in the literature that have looked at the prevalence of nonpainful and painful sensations in clinic populations.

These results have several important implications for healthcare providers. It is important for healthcare providers to not only focus on phantom limb pain but also to assess and treat other types of amputation related pain, in particular the residual limb and back. Using multidimensional assessment measures may provide important clinical information about the frequency, intensity, and bothersomeness of pain (see the chapter by Jensen and Karoly6 for a description of various pain measures). As the data suggest pain may persist for years, healthcare professionals must be careful not to foster unrealistic expectations in their patients by informing them their pain will resolve with time.

Future research efforts should be focused in several areas. First, research on the epidemiology and etiology of back pain after lower extremity amputation is needed. One hypothesis that could be tested is that back pain may result from gait mechanics developed to accommodate a prosthesis. If true, care should be taken to review gait, modifying the prosthesis as necessary, and training or retraining persons with amputations to minimize gait patterns that may produce back pain. Second, similar research is needed that examines the prevalence of other potential locations for pain (hip, knee, shoulder) that may be related to having a lower extremity amputation. Third, residual limb pain warrants attention, given its prevalence and potential interference with prosthetic fitting and use.2 Fourth, to measure pain accurately, researchers should collect multidimensional pain assessments that include all components of the pain experience, including the duration, temporal pattern, and pain affect.6 Psychometrically validated multidimensional measures of pain exist and may offer improved measurement to amputation pain research. Fifth, efforts to examine these pain dimensions and their impact on functional activities and health status also are needed. Finally, given the dearth of evidenced based treatment and the variability of medications used for amputation related pain, controlled clinical trials are needed to evaluate the efficacy of theory driven treatments of amputation related pain.

Back to Top | Article Outline


The authors thank Linda Laing and Christy Vath for their hard work in data collection and entry; and Linda Laing, Dee Malchow MN, RN, and Jerrie Larsen, RN, for their contributions to the design of the Prosthesis Evaluation Questionnaire.

Back to Top | Article Outline


1. Andersson GBJ, Pope MH, Frymoyer JW, Snook S: Epidemiology and Cost. In Pope MH, Andersson GBJ, Frymoyer JW, Chaffin DB (eds). Occupational Low Back Pain: Assessment, Treatment and Prevention. St Louis, Mosby-Year Book 95-113, 1991.
2. Davis R: Phantom sensation, phantom pain, and stump pain. Arch Phys Med Rehabil 74:79-91, 1993.
3. Dillman DA: Mail and Telephone Surveys: The Total Design Method. New York, Wiley 1978.
4. Gift AG: Visual analogue scales: Measurement of subjective phenomena. Nurs Res 38:286-288, 1989.
5. Harwood DD, Hanumanthu S, Stoudemire A: Pathophysiology and management of phantom limb pain. Gen Hosp Psychiatry 14:107-118, 1992.
6. Jensen MP, Karoly P: Self-Report Scales and Procedures for Assessing Pain in Adults. In Turk DC, Melzack R (eds). Handbook of Pain Assessment. New York, Guilford Press 135-151, 1992.
7. Jensen MP, Turner LR, Turner JA, Romano JM: The use of multiple item scales for pain intensity measurement in chronic pain patients. Pain 67:35-40, 1996.
8. Jensen T, Krebs B, Nielsen J, Rasmussen P: Phantom limb, phantom pain, and stump pain in amputees during the first six months following limb amputation. Pain 17:243-256, 1983.
9. Jensen T, Krebs B, Nielsen J, Rasmussen P: Immediate and long-term phantom limb pain in amputees: Incidence, clinical characteristics, and relationship to pre-amputation pain. Pain 21:267-278, 1985.
10. Jensen T, Rasmussen P: Phantom Pain and Other Phenomena After Amputation. In Wall P, Melzack R (eds). The Textbook of Pain. Ed 3. London, Churchill Livingstone 651-665, 1994.
11. Krebs B, Jensen TS, Kroner K, Nielsen J, Jorgensen HS: Phantom Limb Phenomena in Amputees Seven Years After Limb Amputation. In Fields HL, Dubner R, Cervero R (eds). Advances in Pain Research and Therapy. Vol 9. New York, Raven Press 425-429, 1985.
12. Legro MW, Reiber GD, Smith DG, et al: Prosthesis Evaluation Questionnaire for Persons With Lower Limb Amputations: Assessing Prosthesis Related Quality of Life. Arch Phys Rehabil 79:931-938, 1998.
13. Loeser J: Pain After Amputation: Phantom Limb and Stump Pain. In Bonica J (ed). The Management of Pain. Ed 3. London, Churchill Livingstone 244-256, 1994.
14. Melzack R: Phantom limbs. Sci Am 266:120-126. 1992.
    15. Price DD, Harkins SW: Combined use of experimental pain and visual analogue scales in providing standardized measurement of clinical pain. Clin J Pain 3:1-8, 1987.
    16. Sherman R: Stump and phantom limb pain. Neurol Clin 7:259-264, 1989.
      17. Sherman RA, Sherman CJ: Prevalence and characteristics of chronic phantom limb pain among American veterans. Am J Phys Med 62:227-238, 1983.
      18. Stannard C: Phantom limb pain. Br J Hosp Med 50:583-587, 1993.
      19. Syrjala KA, Chapman CR: Measurement and clinical pain: A review and integration of research findings. Adv Pain Res 7:71-101, 1984.
      © 1999 Lippincott Williams & Wilkins, Inc.