Acupuncture for Managing Phantom-Limb Syndrome: A Systematic Review

Overview

A systematic review of the literature was undertaken using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, ¹⁸ to determine the effectiveness of acupuncture for treating phantom-limb syndrome. This review comprised three phases: phase 1 involved a comprehensive database search; phase 2 comprised dual screening of studies by 2 independent reviewers (S.M.M. and G.A.K.); and phase 3 involved the use of validated tools to establish internal validity of included studies by 2 independent researchers (S.M.M. and G.A.K.), as well as quantifying the strength of evidence. Results were then categorized and discussed, using a thematic approach as described in a previous study. ¹⁹

Phase 1: Database Search

A list of key words was drawn up for the terms acupuncture and phantom limb syndrome (see Box 1) with the assistance of a medical librarian and these key words were entered into the PubMed, Embase, PsycINFO, Web of Knowledge, Collection of Nursing and Allied Health Literature,^® Pedro, and Science Direct databases in February 2012. A hand search of the reference lists of each of the identified articles was carried out to identify any suitable studies that were not found during database searching. Results were imported into the Endnote referencing program, ²⁰ which removed duplicates. Reference lists were inspected for any further relevant articles.

Phase 2: Screening of Literature and Data Extraction

The titles, abstracts and keywords were reviewed by 2 independent assessors (S.M.M., G.A.K.) to determine if these items fulfilled predefined inclusion criteria. Inclusion criteria for this review were that studies had to assess subjects who had PLP or PLS post amputation. Amputation had to involve partial or full removal of one or more of the four limbs and congenital limb deficiency was not included, as the mechanisms for this condition are not understood fully. ¹¹ Treatment had to include acupuncture needling (with or without laser and moxibustion) and the effects of treatment on symptoms, including pain, had to be assessed pre- and post-intervention. All types of studies were included, with no limitation on year. Only English-language studies were included.

The 2 researchers met to discuss which articles should be included in the final review, and a consensus was reached (Fig. 1). The full articles of the studies that had been thought suitable were then obtained, with data extracted using a detailed proforma to assess suitability for inclusion. These were again independently reviewed by 2 assessors (S.M.M., G.A.K.), with any studies that did not fulfill inclusion criteria being excluded. From this process, a final list of articles for review was assembled. As this review included studies of mixed methodologies, meta-analysis was not possible and effect size could not be examined. Results were examined using a narrative analysis, and grouped according to outcomes of interest.

OPEN IN VIEWER

FIG. 1.

Flowchart of study inclusion.

Phase 3: Quality Assessment and Strength of Evidence

A “gold standard” assessment tool to assess nonrandomized control trials (non-RCTs) is currently not available. Therefore, two appraisal tools were used to classify the internal validity and quality of the included studies. The Effective Public Health Practice Project (EPHPP) quality-assessment tool ²¹ was used to analyze quantitative studies, while the Critical Appraisal Skills Programme (CASP) quality assessment tool ²² was used to analyze qualitative studies. The internal validity and quality of the included studies was assessed independently by 2 researchers (S.M.M. and G.A.K.), with ratings discussed until a consensus was reached.

The content and construct validity and inter-rater reliability of the EPHPP have met acceptable standards. ²³ This assessment tool has been identified as one of the most appropriate for assessing RCTs and non-RCTs. ²⁴ It has previously been used in systematic reviews with similar methodologies to this one.^25,26 The tool is used to rate six aspects of each study—(1) selection bias, (2) study design, (3) confounders, (4) blinding, (5) data collection methods, and (6) withdrawals/dropouts—as “weak,” “moderate,” or “strong.” Overall quality is classified as “strong” if all sections have strong ratings and no weak ratings; “moderate” if there are less than four “strong” ratings and one “weak” rating; and “weak” if there are two or more “weak” ratings.

The CASP tool ²² is used to assess qualitative research with ten questions that address the rigor of the research methods, the credibility of the findings, and the relevance of the findings. Each question requires a “yes” or “no” answer. The CASP has previously been used in systematic reviews with similar methodologies to the current review.^27,28 Quality of a study is classified as “strong” if more than two-thirds of the answers are “yes”; moderate if between four and six “yes” answers are given; and weak if more than han two-thirds of the answers are “no.” ²⁷

The strength of evidence for the current review was graded according to the Agency for Health Care Policy and Research (now known as the Agency for Healthcare Research and Quality) guidelines. ²⁹ The original rating scale categorized evidence on a 4-point scale (A to D). ³⁰ For the current review, a previously modified version was used that involved a 5-point scale. ²⁷ With this scale, quality of evidence ranges from the highest quality (Level A) to the lowest quality (Level E).

Literature Search

From a total of 6733 database results, sixteen studies fulfilled the inclusion criteria for the current review: ten single-case studies^3,17,31–38; three cross-sectional studies^39–41 ; two case-study series^16,42; and one non-RCT ⁴³ (Table 1). Two case-study series reported treating subjects who had a variety of conditions with acupuncture; in these cases, information was extracted in relation to the description of treatment of the subject with phantom-limb syndrome and, therefore, these two series were treated as single-case studies in the current review.^3,35 Researchers in one study ¹⁷ performed one session of laser treatment without needling and one session of acupuncture needling; for the purpose of this review, only the needling treatment was included. In the three cross-sectional studies, information on the efficacy of acupuncture treatment has been extracted from a larger set of results with information from only the participants who underwent acupuncture therapy being used.

Quality of Evidence

Eleven of the included studies were classified as quantitative in nature, while five were qualitative. Studies that combined the two types of methodology (i.e., quantitative and qualitative) were considered as quantitative for this review, given that the primary outcome measure in each case was quantitative. This was with the exception of one study, ³¹ which did not provide enough information for the study to be classed as quantitative (the quantitative measure was performed pre- but not post-intervention). Of the quantitative studies, all eleven were assessed as being of weak quality when analyzed using the quality EPHPP assessment tool (Table 2). Of the qualitative studies, four were assessed to be of weak quality,^3,31,34,38 and one was categorized as having strong quality, ⁴¹ using the CASP quality assessment tool (Table 3).

Six studies dealt with upper-limb amputees only.^{31,33–36,38} Six studies also dealt exclusively with lower-limb amputees.^{3,16,32,37,39,42} Three studies^40,41,43 dealt with a mixture of upper- and lower-limb amputees, while one study ¹⁷ involved treating a subject with bilateral upper- and lower-limb amputations. Given that one study ⁴⁰ surveyed health care providers instead of patients; no information was available on types of amputations included in that study and, so that study is absent from the following listings. Sites of amputations seen were transtibial,^16,36,37,42 knee disarticulation, ³⁶ transfemoral,^16,17,36,42 hemipelvectomy,^16,36 transpelvic, ³² hip disarticulation,^17,36 two thirds of thumb, ³⁴ four digits of hand, ³⁵ wrist disarticulation, ³⁶ transradial,^17,36,38 transhumeral,^17,36,43 shoulder disarticulation^33,36 and interscapulothoracic. ⁴³ Exact sites of amputation were not stated by three studies.^3,31,41 In one study, ³⁹ the researchers stated that the most common amputation seen was transtibial, but they did not elaborate on the range of amputations.

Amputations were classified as traumatic if the initial injury was traumatic, regardless of subsequent revisions. Nine studies had subjects with exclusively traumatic injuries.^{3,17,31,33–38} One study considered exclusively medical amputations. ³² Four studies had subjects with a mixture of traumatic and medical amputations.^16,39,41,42 The researchers of one study ⁴³ stated that the majority of amputations were needed because the patients had sustained industrial accidents and road-traffic accidents (RTAs); however it was not stated whether medical amputations were also included. The most common cause of amputation was an industrial accident.^{31,35,36,38,42,43} There were also amputations required because of cancer,^16,32,42 military injuries,^3,17,41,42 RTAs,^16,33,43 boating accidents, ³⁷ embolisms, ⁴² falls from a height, ⁴² and congenital talipes. ¹⁶ The researchers of two studies stated only the most common causes among their subjects: trauma; industrial accidents; and RTAs.^39,43 Sherman at al. ⁴¹ found that the majority of their respondents had military-related amputations. The remaining respondents were described as having amputations because of disease, but types of diseases were not specified. Leung and Spoerel ³⁴ reported their participant's amputation as traumatic; however, the exact cause of this amputation was not stated clearly.

Time Since Amputation

Four studies dealt with subjects in the acute stage (i.e., within 1 month of amputation).^{32,36, 37,43} Three studies included subjects in the subacute stage (i.e., between 1 and 6 months of amputation.^16,17,43 Eight studies included patients with chronic symptoms (i.e., ≥6 months.^{16,33–35,38–40,42} The length of time since amputation was not documented by two studies.^31,34 Hao and Hao ³ stated that it was several months since their participant's amputation but did not state how long the participant had the symptoms. Sherman et al. ⁴¹ stated that the average number of years since amputation was 26; however, the range was not reported and it was stated that many respondents had had amputation in the last year or so.

Previous Treatment

It was not possible to extract information about previous treatment from the three cross-sectional studies.^39–41 The researchers of four studies did not state whether their subjects had received other treatments prior to acupuncture therapy.^35–37,42 The researchers of the remaining nine studies described various treatment modalities that had failed to alleviate symptoms prior to acupuncture intervention, including such treatments as pharmacological management, nerve blocks, physical therapy, and neuromal excisions. Eight studies mentioned unsuccessful pharmacological management.^{3,16,17,31–33,38,43} Researchers of four studies described subjects receiving nerve blocks that did not alleviate symptoms.^31,33,34,38 Three studies found that physical therapy did not relieve symptoms.^17,31,43 Two studies reported unsuccessful neuromal excisions.^34,38 Other unsuccessful interventions included a dorsal-column stimulator, ¹⁶ massage, ¹⁷ hypnotherapy, ³⁸ transcutaneous electrical nerve stimulation (TENS), ultrasound, and dorsal-root entry lesion. ³³

Number of Treatment Sessions with Acupuncture

Five studies involved treating 1 or all subjects for one treatment session using acupuncture.^{3,16,17,32,34} Three studies involved treated subjects for four sessions with acupuncture.^16,33,37 Other studies involved completing six ³⁵ and seven ³⁶ sessions of acupuncture. Aldrete and Ghaly ³¹ treated a subject over a 6-month period; however, the number of sessions applied was not stated. Similarly Monga and Jaksic ³⁸ treated a subject with acupuncture over a 3-month period but did not state the number of sessions that were applied. Liaw et al. ⁴³ treated subjects for between one and seven sessions; some subjects discontinued treatment when they were satisfied with the level of improvement experienced. Xing ⁴² treated patients once per day for 3–4 weeks (twenty-one to twenty-eight sessions). It was not possible to extract data about the number of treatment sessions from the three cross-sectional studies.^39–41

Sites of Acupuncture Treatment

Researchers of five studies used points on the contralateral intact limb corresponding to sites of PLP or PLS on the absent limb.^{16,31,36,37,43} Of these five studies, only Bradbrook ¹⁶ and Liaw et al. ³⁶ used recognized acupuncture points. Lu ³⁷ treated a subject with percutaneous electrical nerve stimulation at the thoracic, lumbar, and sacral spine, bilateral knees and the contralateral intact foot. Johnson et al. ³³ used a specific point on the cervical spine, which was found to relieve symptoms; however, this was not a recognized acupuncture point. Monga and Jaksic ³⁸ used Large Intestine, Pericardium, Lung, and Heart points on the contralateral intact limb to treat a transradial amputation and Large Intestine points on the affected side; however, these researchers did not justify their selection of points. Xing ⁴² treated a series of transfemoral and transtibial amputees using Yaojiaji, Gall Bladder, and Bladder points on the affected side; Gall Bladder, Spleen, Stomach, and Liver points on the contralateral intact limb; bilateral Pericardium, Heart, Gall Bladder, and Governing Vessel points, along with the sensory area of the scalp; however this selection of points was not justified. Researchers of three studies^16,32,36 chose points on the basis of their analgesic properties. Fulton ³² and Liaw et al. ³⁶ also chose points for their relaxation-inducing properties. Scalp acupuncture was used by Hao and Hao ³ ; however, exact points were not stated. Auricular acupuncture was utilized by Jacobs and Niemtzow, ¹⁷ who placed one needle at the ipsilateral cingulate gyrus, and by Leung and Spoerel, ³⁴ who inserted a needle into the ipsilateral locus point corresponding to thumb pain. Levine et al. ³⁵ did not state what points were used. Levine et al. ³⁵ and Xing ⁴² also used moxibustion in their treatments. Three studies^35,37,42 reported the use of electroacupuncture. Monga and Jaksic ³⁸ manually stimulated needles. Bradbrook ¹⁶ stated that the needles were stimulated but did not state how this was accomplished. Details of acupuncture treatment were not documented in the three cross-sectional studies.^39–41

Long-Term Follow-Up

Six of sixteen studies followed up with subjects after treatment had finished. Aldrete and Ghaly ³¹ reported that, 18 months after treatment, their subject was pain free and receiving acupuncture for occasional aggravations. Bradbrook ¹⁶ stated that two of three participants remained symptom-free for the remainder of prosthetic rehabilitation. Johnson et al. ³³ reported that their subject was still self-administering acupuncture with pain relief lasting only while the needles were in situ. Leung and Spoerel ³⁴ described their subject at a 3-month follow up as continuing to have good hand function, despite some stump sensitivity. Lu's ³⁷ subject was reported as having returned to work full-time on crutches and that this subject resumed an almost normal physical and mental life. Monga and Jaksic ³⁸ stated that their subject was using TENS once per week to control pain (TENS was used to replace acupuncture so that subject could self-administer treatment) and no longer required pain medication.

Outcome Measures

All sixteen studies used subjective measures, while seven^{16,31,32,34,37,38,40} used a mixture of subjective and objective measures. All sixteen studies used subjects' self-reporting of symptoms. Four studies used a visual analogue scale (VAS).^16,31,33,36 However Aldrete and Ghaly ³¹ only used the measure before treatment and not afterward. Fulton ³² used a numerical rating scale (NRS). Five studies used qualitative descriptors,^{3,16,32,34,37} for example, “telescoping” (the feeling of absent limbs retracting into the body), reduced feelings of heaviness, increasing feelings of warmth, cessation of entire pain syndrome, and normalization of sensation. Four studies^31,33,35,38 used duration of pain relief after treatment as an outcome measure. Hanley et al. ³⁹ used a 5-point anchored scale of treatment success/helpfulness ranging from 1 (“not at all helpful”) to 5 (“extremely helpful”). Sherman et al. ⁴⁰ measured rate of success on an 11-point anchored scale of 0–10, ranging from 0 (“no success”) to 10 (“entirely successful”) and also a score generated by formula (sum of respondents calculated successes divided by number of respondents using treatment). Sherman et al. ⁴¹ used a qualitative scale ranging from “no effect” to “cure.” Jacobs and Niemtzow ¹⁷ used a 0–10 intensity scale. Levine et al. ³⁵ asked subjects to rate percentage of pain (0%=“no pain” to 100%=“worst pain ever experienced”). Liaw et al. ⁴³ used a verbal numerical scale of 0–10 (0=“no pain” to 10 “pre-acupuncture PLP level”). Two studies^37,42 used percentage of pain resolution. Objective measures were not used to measure pain directly but were used to concentrate on changes in functional capacity and analgesic usage. Five studies used functional measures (e.g., ability to tolerate prosthesis),^{16,32,34,37,38} while three studies analyzed the impact on analgesic usage.^31,32,38

Effectiveness of Acupuncture Treatment for Phantom-Limb Syndrome

Results were reported on using three common categories: (1) impact of acupuncture on PLP and/or PLS; (2) impact of acupuncture on functional capacity or mobility; and (3) impact of acupuncture on use of analgesics. All sixteen studies looked at the impact of acupuncture on symptoms of PLP and/or PLS. Five studies looked at the impact on subjects' functional capacity or mobility.^{16,32,34,37,38} Three studies looked at the impact of the intervention on analgesic use.^31,32,38 Treatment was shown to be effective in patients with acute (<6 months, n=6^{16,17,32,36,37,43}) and chronic (≥6 months, n=7^{16,33,34,38,40–42}) symptoms (see Box 2).

Synopsis of Results

This review examined the effectiveness of acupuncture as a treatment for phantom-limb syndrome. Results were based on sixteen studies that included both quantitative and qualitative methodologies. Overall, there was notable evidence from fourteen studies that acupuncture reduces symptoms of PLP and PLS in phantom-limb syndrome. Single-case studies, case-study series, cross-sectional studies, and a non-RCT have used greatly differing forms of acupuncture therapy on subjects of different ages, amputation types, and levels of chronicity, and have shown positive results.

Impact of Acupuncture on PLP and/or PLS

Fourteen of sixteen studies demonstrated that acupuncture reduced PLP and/or PLS. This is consistent with recent research that has shown acupuncture to be an effective treatment for neuropathic pain, ⁴⁴ arthritic pain, ⁴⁵ chronic pain, ⁴⁶ and migraine. ⁴⁷ Treatment was shown to be effective in patients with acute^{16,17,32,36,37,43} and chronic^{16,33,34,38,40–42} symptoms, although it was shown to be more effective within 1 week of the onset of symptoms. ⁴³ Benefits of acupuncture treatment were short-term in many cases. Similar to present findings in phantom-limb syndrome, a Cochrane review found that, for chronic low-back pain, acupuncture is effective for pain relief and functional improvement immediately after treatment and in the short-term only. ⁴⁸ This may however be the result of inadequate doses of acupuncture being provided in some studies because of a lack of clear guidelines in this area. ⁴⁹

Impact of Acupuncture on Functional Capacity and Mobility

Five studies showed that acupuncture had a positive effect on functional aspects of phantom-limb syndrome, such as ability to tolerate prostheses and mobilize.^{16,32,34,37,38} Acupuncture therapy has provoked feelings of “telescoping” (the perception that the phantom limb is progressively shortening in a proximal direction), eventually resolving up to the level of amputation. ⁵⁰ This enabled tolerance of prosthetic limbs and thus functional capacity. ¹⁶ Acupuncture has previously been shown to improve mobility as well as reducing fatigue, depression, and sleeplessness in subjects with rheumatoid arthritis. ⁵¹ Acupuncture therapy has also been effective for improving function in subjects with chronic lower-back pain (cLBP)^52,53 and stroke. ⁵⁴

Impact of Acupuncture on Pharmacological Analgesic Intake

A Cochrane systematic review has suggested that acupuncture is possibly more effective than prophylactic drug treatment in providing analgesia for migraine headaches and has fewer adverse effects. ⁵⁵ Acupuncture has been shown to be an alternative to pharmacological management for phantom-limb syndrome, without the associated side-effects of pharmacological agents, such as nausea and emesis ³¹ as well as memory impairment. ³³ Some forms of pharmacological management control only PLP and not PLS, which is equally functionally restrictive, ³² while acupuncture targets both aspects of phantom-limb syndrome. Acupuncture has previously been shown to reduce levels of pharmacological analgesic use in patients with postoperative pain ⁵⁶ and cLBP. ⁵⁷

Study Type and Quality of Evidence

The type of studies conducted is a serious limitation in this area of research with the presence of ten single case studies and two case study series in this review. A published hierarchy of evidence suggests that the least likely studies to produce good evidence for practice are single-case studies. ⁵⁸ The objective of a case study is to utilize recorded observations and narrative descriptions of an individual's characteristics and responses to treatment to provide a basis for making plausible inferences and generating hypotheses for future research.^59,60 There are advantages to the case-study design in that it offers a rationale for the use of specific rehabilitation interventions or unique applications of treatment strategies and may offer suggestions for treatment modification or more efficient approaches. ⁶¹ Because of its exploratory emphasis, the case study is suitable for generating hypotheses. However, while a case study may include a report of objective findings, it does not provide very strong evidence in support of rehabilitation interventions because of the lack of controlled comparisons (e.g., between treatment and no-treatment conditions), making it difficult to generalize the results to other patients. ⁶¹ Yet, there is no doubt that case reports can provide useful information for further case series, reviews, and original research studies. ⁶²

The quality of studies in this area is another serious limitation, with fifteen of sixteen rated as “weak.” Therefore, the evidence can only be categorized as Level C; although the evidence is generally consistent, it arises from poor-quality studies. Most studies were poorly controlled with significant risks of bias, with data analysis and examination of bias being very poor throughout. Although an acupuncturist administering a real or sham treatment cannot be blinded to the study condition, the potential for bias could be minimized if the acupuncturist were to have the least possible communication with each patient and if the evaluation were to be performed by an independent party. ⁶³ Reporting quality was poor throughout many of the case studies, often with a deficiency of information on selection of subjects, potential influence of concurrent treatments, interventions performed, outcome measures, and follow-ups post treatment. These findings concur with research concerning the quality of acupuncture research, ⁶⁴ which found that quality in reporting details of acupuncture interventions had not improved following the introduction of recommendations specific to this. ⁶⁵

Types of Amputation, Time Since Amputation, and Previous Treatment

Reporting on the types of amputation and times since amputation was not adequate in some studies, limiting the value of their results in the context of current research. However, acupuncture therapy was shown to be effective in both upper- and lower-limb amputations, and in traumatic and medical amputations. Acute and chronic symptoms were treated successfully with acupuncture therapy, with almost all studies featuring subjects who had undergone unsuccessful treatment prior to acupuncture therapy. This may have had an impact upon the success of acupuncture therapy. The timing of treatment for musculoskeletal pain has been reported to be crucial, with early intervention often preventing chronic problems ⁶⁶ and, therefore, it may be important not to wait for conventional treatments to fail before attempting acupuncture therapy. It was not explicitly stated in most studies if treatment was administered by a trained medical professional. It has been previously stated that reporting of the qualifications and experience of practitioners involved in acupuncture case reports is notoriously poor. ⁶⁷ Guidelines have been formulated in order to improve reporting in acupuncture case reports ⁶⁸ and should be followed.

Number of Treatment Sessions

None of the studies contained a plan for the number of treatment sessions prior to applying an intervention. Even within the non-RCT ⁴³ and the two case-study series,^16,42 subjects received differing numbers of treatments. Standardization in intervention is a central component of experimental research so that there is uniform manipulation of the independent variable. ⁶⁹ While interventions will be individualized in practice, further research is essential to formulate guidelines for the number of treatment sessions required.

Sites of Acupuncture Treatment

There is a lack of consensus on appropriate acupuncture treatment sites for phantom-limb syndrome, and a lack of agreement on the optimal acupuncture treatment for any particular condition may mean that some patients do not receive the best treatment. ⁴⁹ Sites of acupuncture were poorly reported or not justified in some studies.^3,38,42 This is consistent with evidence that has identified problems in reporting quality of acupuncture case studies. ⁶⁷ The researcher in one study stated that needles were stimulated but did not state the method used to accomplish this. ¹⁶ Many studies did not use recognized acupuncture points, which have been shown to have a superior analgesic effect compared to non-recognized acupuncture points. ⁷⁰ However, these differences are relatively modest, suggesting that factors independent of effects of needling at specific points are important contributors to the therapeutic effects of acupuncture. ⁴⁶ While both recognized acupuncture points and non-recognized acupuncture points been shown to have analgesic effects, further research should compare the two approaches in the context of phantom-limb syndrome to determine which is more effective.

Long-Term Follow-Up

Only six of sixteen studies performed follow-ups with subjects.^{16,31,33,34,37,38} In some studies, researchers treated subjects with only one session and did not follow-up beyond 1 day of intervention.^3,17,32 Long-term follow up is essential to determine the beneficial or adverse effects of treatment over time. ⁷¹ Therefore, it is difficult to form a conclusion on the long-term effects of acupuncture therapy.

Outcome Measures

Most studies in this review focused primarily on pain intensity. While measuring pain intensity is an important part of assessment and helps evaluate the effectiveness of treatments for phantom-limb syndrome, using this one measure may fail to identify other factors, such as reduced quality of sleep, function, and depression. ⁷² As phantom-limb syndrome is a complex condition, researchers dealing with this problem should consider a multifaceted assessment of symptoms, using published guidelines for core outcome measures in the pain assessment area. ⁷³ There are seven core domains recommended by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials for assessment in studies of chronic pain, including pain, physical functioning, and emotional functioning. ⁷³

It is crucial to include a mixture of both subjective and objective outcome measures in the assessment of phantom-limb syndrome as they are designed to measure different aspects of the same variable and give two distinct types of information. ⁷⁴ Nine studies included in this review relied on subjective self-reported outcome measures only, while seven^{16,31,32,34,37,38,40} used a mixture of subjective and objective measures. Objective measures focused on changes in functional capacity and analgesic usage, indirect measures were focused on the pain experience. Given the complex interaction of sensory, affective, and cognitive components of pain, it is difficult for an objective measure to capture the experience of pain in a clinically useful way. ⁷⁵ Physical functioning and performance tests are frequently used as a proxy for objective pain assessment. ⁷⁶ Measures of physical function described by researchers in the studies examined in this review tended to be vague and based on subjects' self-reporting rather than on clinical assessments. For example, Leung and Spoerel ³⁴ stated that their subject “continues to use her hand quite well.” It has been shown that there are only moderate correlations between self-reported activity limitation and corresponding clinician-measured performance tests. ⁷⁷ The Functional Measure for Amputees measures function of lower-limb amputees in terms of prosthetic usage and is recommended for use in routine clinical practice. ⁷⁸ Medication use and return to work were also used as objective measures by studies included in this review^31,32,37,38 and have previously been used as objective outcome measures in a chronic-pain setting. ⁷⁹

The use of subjective measures is justifiable as patients' self-reporting is the “gold standard” of pain assessment. ⁸⁰ However, in this review just five studies used subjective outcome measures that have been shown to be valid and reliable,^{16,31–33,36} and these outcome measures were not always completed both before and after the interventions.^31,36 Similarly, a review ⁸¹ of subjects with chronic pain or fatigue reported that most of the subjective methods featured had problems with reliability and/or validity. The VAS and NRS have been shown to be valid and reliable for assessing pain, ⁸² while the VAS has been reported to be reliable and repeatable for measuring phantom-limb phenomena and changes following acupuncture. ¹⁶ There is no outcome measure specific to PLP, but, as it is commonly classified as neuropathic pain, ¹ validated tools designed to assess neuropathic pain could be utilized.^83,84

Five studies included in this review used qualitative descriptors,^{3,16,32,34,37} giving insights into patients' individual experiences of illness, of which a good understanding is essential for providing care. ⁸⁵ The McGill pain scale ⁸⁶ is a valid and reliable assessment tool that incorporates these qualitative descriptors and assigns a numerical score that can be used to determine effectiveness of interventions. It would, therefore, provide superior-quality evidence in studies like these rather than the use of descriptors alone. Emotional functioning was not addressed in any of the included studies; but this parameter could be assessed with the Beck Depression Inventory ⁸⁷ or the Profile of Mood States. ⁸⁸ Overall, there needs to be increased utilization of valid and reliable measures rather than formulating unreliable scales for the purpose of individual studies.

Ethics

There was no mention of ethical considerations in any of the studies. Ethics help to govern the conduct of practitioners by coordinating their actions and also apply to the aims or goals of research. ⁸⁹ Johnson et al. ³³ reported using the VAS with a patient post brain injury who had from memory impairment and was later discharged with instructions to self-administer acupuncture. A study on use of the VAS with patients who have Alzheimer's disease (and therefore memory impairment) reported that the majority of midstage patients did not understood the purpose of these scales and recommended the continuation of the search for other tools to assess pain affect in this population. ⁹⁰ Informed consent is one of the most important components of ethical experimentation; however, obtained consent has little meaning unless the subject (or guardian) is capable of understanding what is to be undertaken and all potential hazards. ⁹¹

Limitations

This review itself is not without limitations. Analysis of the most effective type of acupuncture intervention was limited because of heterogeneity among patient populations, interventions, and outcome measures. Selection of acupuncture points is often individualized among patients according to their signs and symptoms, ³ making good-quality RCTs difficult to obtain. Owing to the topic under review, observational, quantitative, and qualitative studies were included, as these comprise much of the information regarding acupuncture and phantom-limb syndrome. This made it necessary to adapt a discursive and thematic methodological approach, as meta-analysis and effect sizes can only be analyzed when only RCTs are included. This narrative approach has, however, been used in several reviews of acupuncture therapy.^49,92 It has also been suggested that the inclusion of qualitative research within systematic reviews may provide a more complete reflection of the experience of subjects, and therefore, lead to more appropriate and effective management. ¹⁹

Another limitation of this review is the lack of an appropriate quality-assessment tool for single-case studies and case-study series. Certain sections of the EPHPP ²¹ and CASP ²² concerning groups, blinding, and randomization were not applicable to cross-sectional studies or case studies and, therefore, automatically generated a “weak” rating. Atkins, and Sampson ⁹³ have formulated a checklist for quality assessment of single-case studies from which a valid and reliable quality-assessment tool could be developed. The inclusion of only English-language articles also limits the scope of this review, as many studies investigating acupuncture are in other languages (e.g., Chinese and Russian).

Implications for Practice and Areas for Future Research

The implications for practice are that practitioners treating patients with phantom-limb syndrome should be aware of acupuncture as a treatment modality and should also be aware that, while it is successful in patients with both chronic and acute symptoms, it is most effective within 1 week of onset ⁴³ and should be not looked on as a last resort when conventional treatment fails.

Further better-quality research must be carried out to determine the form of acupuncture that is most effective, along with optimal treatment times and sites. Quality RCTs using a combination of validated reliable objective and subjective outcome measures are essential to provide evidence for the use of acupuncture therapy in this area. This is in line with the findings of Birch et al. ⁹⁴ who reported that all examined reviews agreed that the methodological rigor of acupuncture clinical trials has generally been poor and that higher-quality clinical trials are necessary. This call to action from Birch et al. ⁹⁴ in 2004 appears to have been largely unheeded in the context of research into acupuncture and phantom-limb syndrome. Greater adherence to STRICTA guidelines ⁶⁴ is essential to improve the standard of research in acupuncture and to develop a mechanism in which guidelines for best practice with regard to the management of phantom-limb syndrome can be documented and shared.

It is also essential to examine the long-term effects of the intervention, as most studies have focused only on the short-term resolution of symptoms. To facilitate clinical and policy decision making, evidence on the long-term consequences of the outcomes used in trials is necessary. ⁹⁵