Abstract
Purpose. Pain control is an ongoing challenge in the oncology setting. Prior to implementing a large randomized trial at our institution, we investigated the feasibility, safety, and initial efficacy of acupuncture for uncontrolled pain among cancer patients. Hypotheses. Our hypotheses were that the acupuncture treatments provided would be (a) feasible, (b) safe, and (c) a beneficial adjunct to pain management. Study Design. This was a single arm, nonrandomized pragmatic pilot study. Methods. Participants experiencing pain ≥4 on a 0 to 10 numeric rating scale received a maximum of 10 treatments on an individualized basis. Recruitment, attrition, compliance, and adverse events (AEs) were assessed. Pain (Brief Pain Inventory–Short Form), quality of life (MD Anderson Symptom Inventory [MDASI]), and patient satisfaction were assessed at baseline and at the end of treatment. Results. Of 115 patients screened, 52 (45%) were eligible and agreed to participate. Eleven (21%) were lost to follow-up, leaving 41 who completed all study procedures. No AEs were reported. Mean pain severity was 6.0 ± 1.3 at baseline and 3.8 ± 2.0 at follow-up (P < .0001). Pain interference was 6.2 ± 2.3 at baseline and 4.3 ± 2.8 at follow-up (P < .0011). On the MDASI, the mean symptom severity was 4.6 ± 1.8 at baseline and 3.2 ± 1.9 at follow-up (P < .0001), and mean symptom interference was 5.8 ± 2.4 at baseline and 4.1 ± 2.9 at follow-up (P < .002). Prescribed pain medications decreased across the course of the study. Patient satisfaction was high: 87% reported that their expectations were met “very well” or “extremely well”; 90% said they were likely to participate again; 95% said they were likely to recommend acupuncture to others; and 90% reported they found the service to be “useful” or “very useful.” Conclusions. Acupuncture was feasible, safe, and a helpful treatment adjunct for cancer patients experiencing uncontrolled pain in this study. Randomized placebo-controlled trials are needed to confirm these results.
Introduction
Pain is a common symptom among cancer patients and can be caused by disease or side effects of treatment. 1 Pain control is an ongoing challenge in the oncology setting and has important implications for patient’s emotional, social, and physical well-being. For many, traditional pain management approaches such as opioid treatment are not an optimal choice, 2 and studies have shown acupuncture has fewer side effects than pharmacological treatments.3-8
Several systematic reviews and meta-analyses9-12 evaluating acupuncture for pain management in cancer patients have been published. Three of these studies9-11 concluded that the evidence was insufficient to support the use of acupuncture because of methodological limitations in the randomized controlled trials that were included. Interestingly, a recent meta-analysis by Choi et al 12 reported that, although acupuncture was not more effective than drug therapy, acupuncture plus drug therapy was significantly more effective than drug therapy alone (n = 437; risk ratio, 1.36; 95% confidence interval = 1.13-1.64; P = .003), implying there is a synergistic effect of acupuncture when added to conventional care.
Although not involving cancer patients, robust evidence supporting the use of acupuncture for multiple chronic pain conditions was recently published bv Vickers et al. 13 An analysis of patient-level data from 29 high-quality randomized controlled trials including 17 922 patients found acupuncture was superior to both sham and no-acupuncture for back pain, neck pain, shoulder pain, osteoarthritis, and chronic headaches (P < .001 for all comparisons). In this study, one of the largest ever conducted on acupuncture, the investigators found modest, but statistically significant, effect size differences between acupuncture and sham (shoulder pain 0.62, spinal pain 0.37, osteoarthritis 0.26, and chronic headache 0.15) and considerably larger effect size differences between acupuncture and no-acupuncture controls (spinal pain 0.55, osteoarthritis 0.57, and chronic headache 0.42). No trials evaluating shoulder pain compared acupuncture to a no-acupuncture treatment group. This study provides strong evidence that acupuncture for chronic pain management is more than a placebo.
Functional magnetic resonance imaging studies also suggest stimulation of acupuncture points can initiate multiple pathways of analgesia through the modulation of neurotransmitter and pain control systems at various levels of the central nervous system (CNS). Huang et al 14 characterized acupuncture stimuli in a recent systematic review and meta-analysis of functional magnetic resonance imaging data and reported the response to acupuncture involves a broad network of brain regions consistent with affective, cognitive, and somatosensory processing. Specific areas found to be involved were the somatosensory cortices, limbic system, basal ganglia, brain stem, and cerebellum. Differences in CNS activity between active and sham acupuncture were also noted.
Electroacupuncture, adding an electrical current between two needles, is often used for pain management, and according to Han et al, 15 specific frequencies of electrical stimulation induce the gene expression of specific neuropeptides in the CNS. The release of opioid peptides evoked by electroacupuncture depends on the frequency of current delivered. 16 For example, a frequency of 2 Hz induces the gene expression of endorphins in the diencephalons, 2 to 15 Hz causes the release of endorphins and enkephalin in the brain and dynorphin in the spinal cord, and a frequency of 100 Hz causes the release of dynorphin in the spinal cord alone. 16 When alternating frequencies are used, these 3 opioids are likely to have a synergistic effect.17,18 Genetic factors may also play a role in individual response to acupuncture, and multiple molecules, in addition to opioid peptides (ie, glutamate [NMDA and AMPA/KA receptors], 5-hydroxytryptamine, and cholecystokinin octapeptide), may help mediate analgesia. 19
Auricular acupuncture, also applied in this study, involves the placement of needles in ear points and is often used for pain management. 20 Alimi et al21,22 evaluated the efficacy of auricular acupuncture for decreasing pain intensity among cancer patients in 2 randomized, placebo-controlled trials and concluded there was a clear benefit.
A recent systematic review of acupuncture in cancer care 23 confirmed that few studies with rigorous scientific methodology have examined its role for pain management in this population. As the overall safety of acupuncture is well established in other populations,3-8 and there are data suggesting a benefit, 24 clinical trials evaluating the use of acupuncture for pain management in cancer patients are warranted. Prior to embarking on a randomized trial at our institution, we conducted a single-arm study to examine the feasibility, initial efficacy, and level of patient satisfaction with acupuncture treatment.
Methods
Participants
Potential participants in this pilot study were identified by faculty in the Pain Management Center at our institution and were referred to the research nurse for assessment of eligibility and to obtain informed consent. Approval from the institutional review board was obtained.
Patients were recruited for participation regardless of gender, age, ethnicity, type of cancer, stage of disease, or previous cancer treatment received. Specific inclusion criteria were (a) ability to understand English, (b) Eastern Cooperative Oncology Group (ECOG) Performance Status of 0 to 2, and (c) pain score ≥4 on a 0 to 10 numerical rating scale (NRS). Patients were excluded if they had any of the following: (a) local infection at or near acupuncture sites used for pain control; (b) deformities that could interfere with accurate acupuncture point location; or (c) known coagulopathy or warfarin or heparin use, including low-molecular-weight heparins. Patients were allowed to participate if they were taking clopidogrel, aspirin, or other nonsteroidal anti-inflammatory agents. Other exclusion criteria included mental incapacitation or significant emotional or psychological disorder that would interfere with the ability to cooperate during this slightly invasive procedure; planned changes in the pain medication regimen, and pregnancy. Finally, patients were excluded if they had a platelet count <100 × 103/µL or a white blood cell count <3 × 103/µL, which were conservative institutional guidelines in place for acupuncture at the time the study was conducted.
As the purpose of this study was to examine the initial effects of acupuncture on the management of uncontrolled pain and to examine the logistics of running this type of protocol, participants were not excluded if they were receiving concomitant chemotherapy or investigational anticancer therapies; however, eligible patients had to remain on a stable course of pain medication throughout the treatment period, although minor changes and reductions in dosage were permitted. No specific prior pain treatment regimen was required for participation.
Patients’ medical records were reviewed, and the following baseline characteristics were recorded: primary cancer diagnosis; stage of disease at time of study entry; presence or absence of metastasis; history of cancer treatment; pain management history; type and location of pain and its suspected etiology. Prescribed pain medications at the time of study entry and at the end of treatment were also recorded.
Treatment and Procedures
After baseline measures for pain and quality of life were obtained, patients received individualized acupuncture treatments 1 to 3 times per week, depending on the acupuncturists’ recommendation as well as the patient’s scheduling preferences. Two licensed and experienced acupuncturists (more than 30 years combined), who were credentialed by the hospital, provided all treatments. Both practitioners graduated from the same standardized masters’ degree program in Houston, Texas; thus, variability in treatment approach and methods was limited.
Patients received a maximum of 10 treatments in the acupuncture clinic at our institution. On the first visit, patients who had no contraindications were given the option of undergoing a single treatment consisting of the insertion of gold-plated stainless steel ear studs (Aiguille Semi-Permanent [ASP] auricular acupuncture needles; SEDATELEC, Lyon, France) placed consecutively and bilaterally into the following points: Cingulate gyrus (CG), Shenmen, Point zero, and Subcortex. 25 Patients were eligible for this initial session if they did not have allergies to gold or other dermatologic issues and were capable of safely maintaining the retention needles. They were instructed that the ear studs would remain in place for 3 to 5 days and were told they should remove them with an alcohol swab or tweezers if any evidence of redness or irritation developed. Patient-reported pain scores were recorded on a 0 to 10 NRS before and after this initial treatment.
During subsequent treatment sessions, patients were assisted into a comfortable position and assessed by the acupuncturist to determine the most appropriate treatment approach. Points were selected on an individualized basis according to the theory of traditional Chinese medicine, 26 patient’s medical history, location and type of pain, and concomitant symptoms. All treatments were given using needles (32-40 gauge and 15-40 mm length) manufactured by Seirin Corporation (Shizuoka, Japan). A list of the most common points used, by location, is provided in Table 1.
Most Common Acupuncture Points Used by Location.
The average number of points used per treatment was 14 (range 8-23).
Standardized techniques for point location were used.26,27 As documented in the patient’s electronic medical record, if there was active disease, needles were not placed in the same anatomical region as the tumor. After the skin was prepped with 70% alcohol, the needles were placed at the selected body points and left in place for approximately 25 minutes. Needle insertion endpoints were the standardized recommended depth of insertion26,27 or achievement of de qi sensation. De qi sensation is defined as the sensation of numbness, tingling, or warmth at the needle insertion site. It may also be experienced by the acupuncturist as the needle being “grabbed” by underlying tissues. The recommended depth of insertion26,27 ranged from 0.5 to 1.5 inches for body points and 1 to 2 mm for ear points.
Electrical Stimulation
At each visit, patients were assessed by the acupuncturist to determine whether or not electrical stimulation should be added. No electrical stimulation was applied if the patient had a cardiac pacemaker or a history of prior negative response or intolerance to electrical stimulation. Electrical stimulation was added after de qi sensation was elicited and was applied to acupoints by placing lead wires on the needles connected to an electro-acupuncture stimulator (IC-1107; ITO Co, Ltd, Tokyo, Japan). The electrical stimulation was applied to points along the same channel with the leads attached from negative to positive in the same direction as channel flow. 26 The stimulator was programmed to alternate between high and low frequencies (2-100 Hz) using a continuous (dense-disperse) waveform for 20 minutes.
Measures
Patients completed self-report questionnaires at baseline and after the last acupuncture session. Data collected shortly after the last treatment included assessment of current pain, average pain, worst/least pain within the past 24 hours, and the interference of pain with daily activities and quality of life (QOL) within the past 24 hours.
Brief Pain Inventory–Short Form
The Brief Pain Inventory–Short Form (BPI-SF) is a validated, widely used, self-administered questionnaire to assess the severity of pain and its impact on daily functioning. 28 Item 1 assesses whether or not patients are currently experiencing pain other than minor headache, sprain, or toothache. The second item asks patients to indicate on a whole body diagram the exact location of their pain, and items 3 to 6 measure pain severity on a 0 to 10 numeric scale (worst, least, and average pain in the past 24 hours and current pain level). Items 7 and 8 assess treatments and pain relief from treatments or medications within the past 24 hours. Finally, item 9 includes 7 subratings that determine the degree to which pain has interfered with the patient’s QOL within the previous 24 hours. Assessment of “pain at its worst in the last 24 hours” (item 3 on the BPI-SF) satisfies the US Food and Drug Administration guidelines for measuring a pain reduction treatment effect, and a reduction of at least 2 points on a 0 to 10 scale indicates a clinically significant improvement. 29 For this study, patients, therefore, had be experiencing pain ≥4/10 on the NRS.
MD Anderson Symptom Inventory
The MD Anderson Symptom Inventory (MDASI) is a brief questionnaire that measures on a scale of 0 to 10 the severity and impact of cancer-related symptoms on the patient’s QOL over the previous 24 hours. 30 Thirteen core items ask about symptoms common to cancer and cancer treatment. An additional 6 items evaluate the extent to which cancer-related symptoms interfere with the patient’s daily activities.
Participant Feedback Form
The Participant Feedback Form (PFF) is a standard departmental form used to assess patient satisfaction with services offered. It consists of 6 questions that evaluate on a scale of 0 to 3 whether or not the service met the patient’s expectations and whether or not they would recommend it to others. An additional 3 items are open-ended and allow patients to make recommendations for future program or service offerings. The PFF was administered to all study participants at the end of the treatment period.
Statistical Considerations
Our criteria for determining study feasibility were that half of the patients screened would be eligible, at least 80% of eligible patients would agree to participate, and we would have a dropout rate ≤25%. Additionally, 3 different methods of analysis were conducted to measure the treatment effect. First, paired t tests were performed to evaluate differences from baseline to the end of treatment for patients who completed all prescribed acupuncture treatment sessions and study procedures. For additional comparison, data were analyzed with simple imputation by replacing missing data with group means at the corresponding time point. Furthermore, intent-to-treat procedures with multiple imputations were used to evaluate all patients who consented to participate and completed at least 1 treatment session using the SAS V9.2 MI procedure with the Markov chain Monte Carlo method to impute 10 times and then using the MIANALYZE procedure to generate statistical inferences. 31
Results
Of the 115 patients screened, 56 (49%) met all eligibility requirements. Of the eligible patients, 52 agreed to participate, with 4 declining either because of scheduling conflicts, transportation problems, or because they stated, “I don’t like needles.” Participant characteristics are shown in Table 2. The average length of time patients had been receiving treatment for pain management was 31 months (range, 0.25-107 months). Most participants (67%) reported pain in multiple locations as follows: back (40%), neck/shoulders (23%), trunk/thorax (15%), lower extremities (15%), upper extremities (13%), abdomen (8%), joints (6%), peripheral neuropathy (4%), rectum/genitals (4%), general body pain (4%), headache (2%), and orofacial pain (2%). All patients received acupuncture 1 to 2 times per week. Thirty-nine (75%) patients received auricular ASP needles on their first visit and electroacupuncture stimulation on subsequent visits.
Patient Characteristics.
NED indicates no evidence of cancer at the time of study entry.
Some patients experienced pain at multiple locations.
CIPN indicates chemotherapy-induced peripheral neuropathy in all 4 extremities.
Patient flow throughout the study is presented in Figure 1. Eleven (21%) patients who entered the study were lost to follow-up owing to difficulty in scheduling treatments, changes in pain medications, or disease progression. For these 11 patients, the average number of treatments received was 5 (range, 2-10) over an average of 4 weeks (range, 1-13). One patient failed to complete the second page of one questionnaire. Therefore, 41 patients completed all study procedures and the prescribed acupuncture treatment regimen (average number of treatments = 8; range, 2-10) over an average of 5 weeks (range, 1-14). Of these, 71% (29/41) received auricular ASP needles on their first visit. On subsequent visits, 71% (29/41) received electro-acupuncture at body points, although not necessarily the same 29 patients who received auricular acupuncture.
Patient flow diagram.
Paired t tests revealed significant differences in symptom severity scores and interference scores on both the BPI-SF and the MDASI between baseline and the end of the study for the 41 evaluable patients (see Figure 2 and Table 3). For the BPI-SF, there were significant reductions in both the pain severity score and the interference score from baseline to follow-up (P < .0001 and P < .0011, respectively). Reduction in “worst pain in the last 24 hours” was 2.4 points (P < .0001). For the MDASI, there were also significant reductions in both the symptom severity (P < .0001) and interference (P < .002) scores from baseline to follow-up. When the analyses were repeated using intent-to-treat procedures with either simple imputation (ie, replacing missing values with group means at the corresponding time point) or multiple imputations, all findings remained highly significant.
Brief Pain Inventory–Short Form and MD Anderson Symptom Inventory scores.
Comparison of Patient Outcomes.
Three different methods of analysis were conducted to measure treatment effect. First, paired t tests were performed to evaluate differences from baseline to the end of treatment for patients who completed all prescribed acupuncture treatment sessions and study procedures.
For additional comparison, data were analyzed with simple imputation (SI) by replacing missing data with group means at the corresponding time point.
Finally, intent-to-treat procedures with multiple imputations (MI) were used to evaluate all patients who consented to participate and completed at least one treatment session using the SAS V9.2 MI procedure with the Markov chain Monte Carlo method to impute 10 times and then using the MIANALYZE procedure to generate statistical inferences. 31
Subset analyses were performed comparing the 29 patients who received both ear and body point acupuncture with electrical stimulation versus the 12 who received body point acupuncture only with no electrical stimulation. No significant differences between the 2 groups were found on either the BPI-SF or MDASI. Additionally, examination of pre- and posttreatment pain scores (0-10 NRS) immediately before and after the auricular acupuncture given on the first visit in this subset of 29 patients revealed significant reductions in pain (mean change in scores: −1.3 ± 1.1; P < .0001).
For 34% (14/41) of patients, prescribed medications for pain were reduced by the end of treatment, and their BPI-SF pain severity scores and interference scores were also significantly reduced compared with baseline (P = .0003 and P = .07, respectively). Among these patients, worst pain (P = .006), least pain (P = .003), average pain (P = .01), and pain “right now” (P = .0002) were also significantly lower at the end of treatment, as were their MDASI symptom severity and interference scores (P = .02 and P = .01, respectively). For 10% (4/41) of patients, prescribed pain medications were increased. Among these patients, only the BPI-SF and MDASI interference scores were significantly lower at follow-up (P = .02 for both).
For 44% (18/41) of patients, the prescribed pain medications were unchanged, yet the BPI-SF pain severity and interference scores were significantly reduced from baseline (P = .0002 and P = .01, respectively). Among these patients, worst pain (P = .007), least pain (P = .003), average pain (P = .002), and pain “right now” (P = .0008) were also significantly lower at the end of treatment, as were MDASI symptom severity and interference scores (P = .004 and P = .007, respectively). For 12% (5/41) of patients, changes in prescribed pain medications were unclear due to missing data. Among these patients, the BPI-SF pain severity and interference scores were significantly reduced from baseline (P = .02 and P = .04, respectively). Although a significant reduction from baseline was found for worst pain (P = .009), least pain, average pain, and pain “right now” were not significantly improved. Likewise, no significant improvement was found for MDASI symptom severity or interference scores.
Changes in pain medication prescription were as follows: 81% (34/42) of patients had opioids prescribed at baseline, with only 52% (22/42) at follow-up; 43% (18/42) had nonopioid analgesics prescribed at baseline, with 31% (13/42) at follow-up; 12% (5/42) had tricyclic antidepressants prescribed at baseline, with 2% (1/41) at follow-up; 29% (12/42) had other antidepressants prescribed at baseline, with 17% (7/42) at follow-up; 52% (22/42) had gabapentin prescribed at baseline, with 40% (17/42) at follow-up; 7% (3/42) had steroids prescribed at baseline, with 5% (2/42) at follow-up; 12% (5/42) had muscle relaxants prescribed at baseline, with 17% (7/42) at follow-up; and 14% (6/42) had benzodiazepines prescribed at baseline, with 0% (0/42) at follow-up. Additionally, 2/42 (5%) patients had electronic pain pumps at baseline, and 1/42 (2%) had a pain pump at follow-up.
In terms of patient satisfaction, 87% stated that the course of acupuncture met their expectations “very well” or “extremely well,” 90% said they were likely to have acupuncture again, 95% said they were likely to recommend acupuncture to others, and 90% said they found the acupuncture to be “useful” or “very useful.”
Discussion
Pain among cancer patients takes many forms and is often difficult to manage. Furthermore, the logistics of providing acupuncture service in the setting of a busy clinic are challenging. This study was, therefore, designed to determine the feasibility, safety, and initial efficacy of acupuncture to treat uncontrolled pain in a cancer patient population in order to inform a future larger randomized trial. Although only 49% of patients screened were eligible for acupuncture therapy, all other feasibility criteria were met: Of eligible patients, 93% agreed to participate, and the dropout rate was 21% during the study period. The treatment was safe and there were no adverse events reported by the patients.
In terms of an initial assessment of effect, there was a highly significant reduction in pain severity and interference scores between baseline and the end of treatment. Similar improvements were seen in overall QOL and interference scores. It is also notable that 78% of patients had either a reduction or no change in their pain medications, yet their pain outcomes and QOL improved. Importantly, the reduction in “worst pain in the last 24 hours” for all patients was 2.4 points, which represents a clinically significant reduction in pain scores. 29
Although changes in pain medications should be interpreted with caution, as patients who took less medication may have experienced more pain and vice versa, 34% of patients had fewer pain medications prescribed at the end of treatment compared with baseline. Furthermore, the percentage of patients who were prescribed opioids was reduced from 81% at baseline to 52% at the end of treatment. Moreover, there were also reductions in the percentages of all other categories of pain medications, except for muscle relaxers, which were prescribed at a slightly higher rate. The reason for this small increase in use of muscle relaxers is unclear; however, it was due to only 2 patients out of 42 and should be interpreted with caution. Overall, most patients reported a benefit from acupuncture treatment, with no adverse events reported.
Even though the study found significant improvement in pain and QOL outcomes, the specific effects of acupuncture needling are not well understood, and it is important to remember that the natural history of pain is to decline over time. The use of acupuncture for pain control in a cancer population also likely has a large placebo effect, 32 which needs to be controlled for in future studies. However, it is important to note that on average these patients had been experiencing pain for more than 2 years and were on pain medications that did not seem to bring them adequate pain relief, and for which there is also a substantial placebo effect. Other limitations of the study were that there was no long-term follow-up; for patient convenience, data were collected shortly after the last treatment session. However, the questionnaires evaluated average symptoms and symptoms over the past 24 hours, and according to the Food and Drug Administration, assessment of “pain at its worst in the last 24 hours” on the BPI-SF is sufficient for measuring a pain reduction treatment effect. 29 An additional limitation was that the acupuncture treatments (ie, point selection as well as number and frequency of treatments) were individualized. Although individualized treatments are representative of actual clinical practice, for research purposes, using a non-fixed treatment approach makes interpretation of the results difficult, as the specific effects of acupuncture remain unknown.
Conclusions
Acupuncture was found to be a feasible, safe, and helpful adjunct for treating uncontrolled pain among cancer patients in this preliminary trial. Randomized placebo-controlled trials are needed to confirm the efficacy of acupuncture for pain management in a cancer population.
Footnotes
Acknowledgements
We would like to thank Dawn Chalaire and the Department of Scientific Publications, The University of Texas MD Anderson Cancer Center for helpful editorial comments on this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support was provided in part by the United States National Cancer Institute (NCI) grant CA121503, the NCI Cancer Center Support Grant CA016672, and The University of Texas MD Anderson Cancer Center.