OUP user menu

Survey of Phantom Limb Pain, Phantom Sensation and Stump Pain in Cambodian and New Zealand Amputees

Kelly Patrick Anthony Byrne MBChB
DOI: http://dx.doi.org/10.1111/j.1526-4637.2011.01105.x 794-798 First published online: 1 May 2011


Objective. The primary objective of this study is to compare the prevalence of phantom limb pain in New Zealand and Cambodian amputees and to assess the demographics of a sample of amputees from these two countries.

Design. All participants were interviewed using a 12-question survey that covered demographic data and reason for amputation and assessed the presence of phantom limb sensation, phantom limb pain, and stump pain. Amputees attending an artificial limb center in Cambodia were approached and interviewed in person. New Zealand amputees attending the Waikato artificial limb center were randomly selected and interviewed by phone.

Results. There was no statistically significant difference in phantom limb sensation, phantom limb pain, or stump pain between the two groups. There was a much higher unemployment rate in the Cambodian amputees. There were no other statistically significant differences between the groups

Conclusion. Despite very different environments, there was no difference in the phantom limb pain between the groups. One possible explanation is that the severity of neurological injury associated with amputation overrides all the other risk factors that influence the development of other chronic pain syndromes.

  • Cambodia
  • New Zealand
  • Amputation
  • Phantom Limb Pain
  • Survey


Phantom limb pain is a complicated disorder affecting 60–80% of amputees [1]. Neurological changes in the periphery, spinal cord, and cerebral cortex are all likely to contribute to phantom limb pain [2].

The primary aim of this study is to compare the prevalence of phantom limb pain between two groups of amputees in New Zealand and Cambodia. The secondary aims are to compare the prevalence of phantom limb sensation and stump pain in these two groups. Additional demographic and employment data for these two groups were also collected.


Ethics committee approval for this study was obtained from the New Zealand Northern Y ethics committee, which has jurisdiction in the Waikato region. There are no functioning local, regional, or national ethics committees in Cambodia.

During a 3-month period from July to August 2006, 29 Cambodian amputees were interviewed at the Veterans International Artificial Limb center at Kien Khleang, a nongovernmental organization run by Cambodians and funded via United States Agency for International Development and by donation. This sample consisted of 29 consecutive patients who presented to the artificial limb center, either for fitting and training in the use of a prosthesis or for follow up and refitting of a currently used prosthesis, a service that is provided free of charge. Verbal consent was obtained from all patients prior to commencing the interview through a locally trained medical doctor. All patients approached agreed to take part in the survey.

The New Zealand group was broadly matched for age to the Cambodian group by generating a list of amputees aged between 20 and 50 years old who attend the Waikato Artificial limb center. This list consisted of 296 people. Each of the patients was allocated a number, and then 80 numbers were randomly selected. These people were then phoned and asked to participate in the study. The first 31 people who were able to be contacted were included in the study. All patients approached agreed to take part in the survey. Two patients were subsequently excluded on the basis that they had congenital amputations as there were no comparable patients in the Cambodian group. The remaining 29 patients were included in the study. All patients were asked for verbal consent prior to undertaking the study and were sent an information sheet about the study giving them the opportunity to withdraw from the study.

Participants from both countries were interviewed by the same interviewer using the same set of 12 questions (included in Appendix), which covered basic demographic data, reason for the amputation, time since amputation, employment status before and after the amputation, questions inquiring about the presence of phantom limb sensation, pain and stump pain, and use of any analgesic agents. Patients rated the current amount of pain in their daily life on an 11-point verbal pain scale, where 0 was no pain and 10 was the worst pain imaginable. There is no standardized or validated questionnaire that has been used in previous studies into phantom limb pain [3–5]; however, there were many similarities between the questionnaire used and those used in previous studies, allowing comparison with previous studies.

Data were collated on an Excel spreadsheet and mean and standard deviation was calculated with the Analyse-it package for Excel (Version 2.05, December 2007, Analyse-it Software Ltd, Leeds, UK). The chi-squared testing was used to compare the categorical data collected and two-tailed unpaired Student's t-test was used to compare the numerical data. The statistical testing was undertaken using the GraphPad QuickCalcs statistics package (GraphPad Software Inc, La Jolla, CA).


The major finding of this survey is that there are no statistically significant differences between the groups in terms of prevalence of phantom limb pain between the amputees in Cambodia and New Zealand. Fifteen (51.7%) in the Cambodian group had phantom limb pain, while 20 (69%) in the New Zealand group had phantom limb pain (P = 0.18) (Table 1).

View this table:
Table 1

Prevalence of phantom limb pain, phantom limb sensation and stump pain

Cambodia (%) [N, SD]New Zealand (%) [N, SD]P value
Phantom limb sensation  23 (79.3)24 (82.2)0.738
Phantom limb pain  15 (51.7)20 (69)0.180
Stump pain  18 (62.1)14 (48.3)0.293
Mean pain scores (phantom pain only) 6.5 [6, 1.8]7.1 [12, 2.5]0.610
Mean pain scores (stump pain only)4.88 [9, 2.8]4.0 [6, 3.4]0.593
Mean pain scores (combined)  5.6 [24, 2.48]6.4 [26, 2.95]0.307

In terms of the secondary outcomes that were considered, there were no statistically significant differences in terms of phantom limb sensation and stump pain in the two groups of amputees. Twenty-three (79.3%) from the Cambodian group experienced any phantom limb sensation, while 24 (82.2%) from the New Zealand group experienced any phantom limb sensation (P = 0.738). Eighteen (62.1%) from the Cambodian group had stump pain, while 14 (48.3%) from the New Zealand group had stump pain (P = 0.293).

Demographic data were similar between the groups (Table 2) as was expected due to the age matching that took place. In both groups, there were 25 men (86.2%) and four women (13.8%). The mean age was 40.3 years in the Cambodian group and 41.7 years in New Zealand group. The mean time since amputation was 10.4 years in the Cambodian group and 15.1 years in the New Zealand group, a difference that approached statistical significance (P = 0.052).

View this table:
Table 2

Demographic characteristics of the two groups

Cambodia (%) [SD]New Zealand (%) [SD]P
Male  25 (86.2)  25 (86.2)  1.0
Female    4 (13.8)    4 (13.8)  1.0
Mean age (years)40.3 [10.5]  41.7 [4.8]  0.516
Time since amputation (years)10.4 [8.2]15.11 [9.8]  0.052
Employed at time of amputation  27 (93.1)    28 (96.6)  1.0
Employed at time of interview    9 (31)  24 (82.8)<0.0001

There were no significant differences in the type of amputation suffered by the interviewees. When upper limb amputations of all types were combined (Table 3), there were more in the New Zealand group, although this did not reach statistical significance (P = 0.053).

View this table:
Table 3

Type of amputation in survey participants

Type of amputationCambodia (%)New Zealand (%)P
BKA18 (60)15 (48.4)0.444
AKA 8 (26.6) 6 (19.4)0.554
Bilateral BKA 1 (3.3) 1 (3.2)1.0
below elbow 1 (3.3) 2 (6.5)1.0
Above elbow 0 4 (13)0.113
Arm and BKA 0 1 (3.2)1.0
Forefoot 1 (3.3) 00.492
Combined lower leg amputation29 (96.7)24 (77.4)0.053
Combined upper limb amputation 1 (3.3) 7 (22.6)0.053
  • BKA = below knee amputation; AKA = above knee amputation.

The reasons for amputation differed between the two groups (Table 4), with 18 (62.1%) of the Cambodian group being involved in land mine accidents. The next most likely reason for amputation in the Cambodian group was cancer and road traffic accidents, which accounted for three patients (10.3%) in each of those categories. In the New Zealand group, 16 (55.2%) patients had amputations due to road traffic accidents, and a further six (20.6%) had amputations due to other accidents, including work accidents involving forestry or mining equipment. The remaining amputees had diverse reasons for amputation.

View this table:
Table 4

Reasons for amputation in the two groups

Reason for amputationCambodia (%)New Zealand (%)
Land mine accident18 (62.1)
Cancer 3 (10.3)
Vehicle accident (car/bike/motorbike) 3 (10.3)
Nonhealing ulcer 1 (3.4)
Chronic foot infection 1 (3.4)
Ischaemic leg 1 (3.4)
Unexploded munitions 1 (3.4)
Severe burn 1 (3.4)
Motorbike accident 8 (27.6)
Car accident8 (27.6)
Other accident (work/hunting)6 (20.6)
Chronic infection3 (10.3)
Cancer2 (6.9)
Amputation of nonfunctional leg (cerebral palsy)1 (3.4)
Landmine1 (3.4)

The number of people employed at time of amputation was 27 (93%) in the Cambodian group and 28 (96%) in the New Zealand group (Table 2). At the time of interview, nine (31%) from the Cambodian group were employed and 24 (83%) from the New Zealand group were employed, a difference that was statistically significant when compared using chi-squared testing (P < 0.001).

Patients described a wide variety of types of pain; however, descriptions for pain consistent with neuropathic pain (burning, stinging, electric shock-like) were common. In the Cambodian group, three (10.3%) patients had pain that interfered with their lives, while in the New Zealand group, seven (24.1%) of the patients reported that pain interfered with their life (Table 5), which was not significantly different (P = 0.297). Eight (27.6%) patients in the Cambodian group were currently taking medications and seven (24.1%) of the patients in the New Zealand group were taking medication (Table 5). Several patients in the New Zealand group volunteered that they had previously been taking medication but had stopped due to either side effects or lack of efficacy.

View this table:
Table 5

Degree of disability and medication use in the two groups

Cambodia (%)New Zealand (%)P
Does pain interfere with life?0.297
  Yes 3 (10.3) 7 (24.1)
  No26 (89.7)22 (75.9)
Current use of medications1.00
  Yes 8 (27.6) 7 (24.1)
  No21 (72.4)21 (75.9)
Type of medication being used0.619
  Paracetamol 5 (62.5) 3 (42.8%)
  Unknown analgesic 3 (37.5)
  Opioid 2 (28.6)
  Gabapentin 1 (14.3)
  Amitryptiline 1 (14.3)


The major finding in this study is that there is no difference in the prevalence of phantom limb sensation, phantom limb pain, and stump pain between the two groups of amputees. Given the greater access of New Zealand amputees to health care and the higher standard of health care in New Zealand, it could be reasonable to expect that phantom limb pain would be less prevalent in the New Zealand amputees. Unfortunately, this study does not go any way to explain why there is no difference in phantom limb pain between these two groups. However, this result is useful in that it seems to indicate that perhaps the cultural, social, and economic environment that an amputee is exposed to may not influence the development of phantom limb pain.

The prevalence of phantom limb sensation, phantom limb pain, and stump pain are broadly in line with previous studies [4–6] given the small sample numbers in all these studies. This supports the accuracy of the data collected. In a previous study of land mine accident survivors in Cambodia and Kurdistan, the Cambodian group had a 48% rate of chronic pain syndromes in a sample of 25 patients [4]. Another comparable study undertaken in Sierra Leone of 40 arm amputees had a phantom sensation rate of 93%, phantom limb pain rate of 33%, and stump pain rate of 100% [6]. An explanation for the relatively low level of phantom limb pain in the Sierra Leone study is suggested by Dijkstra et al. [7], who showed that upper limb amputees had a lower rate of phantom limb pain than lower limb amputees.

Retrospective studies have shown that low education status, pre-amputation pain, and untreated depressive symptoms are more common in patients with phantom limb pain than in patients with amputation but no phantom limb pain [1,8,9]. Education status between the two groups studied here is clearly different, although this was not quantified. It was not investigated whether there were differences in pre-amputation pain or incidence of untreated depression in these two groups. Due to the overwhelming nature of the neurological injury sustained during amputation, it is possible that the risk factors identified in retrospective analysis of amputees play a relatively minor role in the development of phantom limb pain. Possibly in other chronic pain states where the neurological injury is more subtle, these factors play more of a role.

Despite high mean pain scores, the level of interference with life reported by the people was low. Only three patients (10.3%) in the Cambodian group and seven patients (24.1%) in the New Zealand group said that the pain they experienced interfered with their lives. This was not a statistically significant difference between the two groups (P = 0.297). This correlates well with the findings of a previous study where pain-related disability was measured by the Chronic Pain Grade in which 75% of amputees fell into the low disability/low pain intensity or low disability/high pain intensity categories, indicating low levels of inference with activities of daily life [5].

The New Zealand group had a longer time since amputation, which came close to statistical significance (P = 0.052) (Table 2). Some studies suggest that prevalence of phantom limb pain falls with time since amputation, although this has not been a universal finding, and may alter with the way in which the data are collected. Patient's memory of what previous pain was like has been shown to be unreliable [9]. There have been no long-term (greater than 2 years) prospective trials looking at how phantom limb pain evolves over time.

There was no significant difference in employment status at time of amputation; however, there was a large difference in employment status at time of interview, which was statistically significant (P < 0.0001). This can be explained by two reasons. First, 16 (55.2%) of the Cambodian amputees were employed as soldiers at the time of their amputations during a time of conflict in Cambodia. Subsequent to this time, the conflict has largely resolved and large numbers of soldiers have been decommissioned and it is likely that injured soldiers were the first to be discharged from duty. Second, the current economic climate in Cambodia means that large numbers of the general population are unemployed. Lack of governmental infrastructure means there are no reliable figures; however, World Bank figures from 2004 put the poverty rate at 36% (income below U.S. $1 a day), likely reflecting a high unemployment rate [3]. The unemployment rate in New Zealand at the time the patients were being interviewed was less than 5% [10]. Therefore, this difference is more likely to reflect the economic differences between the two countries than differences in disability in the two groups of amputees.

Employment status and phantom limb and residual limb pain has previously been studied in Japanese amputees, and type of pain or degree of disability from pain was not found to be associated with whether or not the patient was employed [11].

The use of analgesic agents was similar between the two groups and low compared with the number of people who reported pain. What was surprising in the New Zealand group, which does not come through in the results shown, is the number of people who had been started on treatment only to discard it due to lack of efficacy and side effects. There were no direct questions about this, but several people volunteered this information. This fairly accurately reflects the current state of evidence-based treatment of phantom limb pain, where numerous treatments have been tested with no great benefits uncovered. As previously summarized in a review by Halbert et al., “there is little evidence from randomized trials to guide clinicians with treatment”[12].

Sources of Error

The sample size in this study is too small to make any unequivocal statements about the differences in phantom limb pain and stump pain in these two populations. A posteriori analysis of the sample size required to show a difference in phantom limb pain reveals that a sample size of 126 is required to show a statistically significant difference, nearly twice the number sampled. For a statistically significant difference to be seen in stump pain, a sample size of 218 would be required.

Pain was sampled at a single point in time and difference results may have been obtained if patient's pain was tracked over time. In addition, the questionnaire used did not thoroughly investigate patient's level of disability or other characteristics that may contribute to chronic pain, such as depressive symptoms.

Differences in understanding, education of the patients, and the translation of the questionnaire into Cambodian may have also affected the accuracy of the data collected.


In conclusion, it appears that in these two communities, there is no difference in phantom limb sensation, phantom limb pain, or stump pain despite exposure to quite different environments. This study shows that the prevalence of these problems is broadly in line with previously published studies and that phantom limb pain is a considerable problem in these two communities.

This study adds to the prevalence data from previously conducted studies and suggests that environmental factors may not have significant influence on the development of phantom limb pain. More extensive research needs to be conducted with larger sample sizes and other communities comparing the prevalence of phantom limb pain. Analysis of the psychosocial factors in different communities and their prevalence of phantom limb pain may lead to development of nondrug strategies for treatment or prevention of phantom limb pain.

Finally, despite relatively high pain scores, overall disability as a result of this seems to be low. This is an area that would benefit from further study, directly aimed at assessing overall disability from limb amputation.


There were no grants or financial support for this project from any source. The author has no financial relationships with any companies, and there are no other potential conflicts of interest present.


I would like to thank the staff of the Veterans International Artificial Limb Centre in Phnom Penh and Claude and the rest of the staff at the Waikato Artificial limb center for making this study possible. Thank you to Dr. Rattana Kruy for translating during interviews in Cambodia. I would also like to thank Dr. Anders Johnson, Dr. John Barnard, and Dr. Malcolm Hogg for proofreading this article and for their useful suggestions and support.



  1. When did your amputation occur?

  2. What was the cause of your amputation?

  3. Did you have a job before your amputation? If so, what was it?

  4. Do you have a job now? If so, what is it?

  5. Do you ever have the sensation that your missing limb is still there?

  6. If yes, does this trouble you at all?

  7. Do you ever have pain in your missing limb?

  8. Do you ever have pain at the site of your amputation?

  9. If 0 was no pain and 10 was the worst pain in the world, where would you rate your pain?

  10. Can you describe the type of pain that you have?

  11. Does this pain ever stop you from doing anything?

  12. Do you take any medications for your pain? Or medicines for other conditions?


View Abstract