Abstract
Objective:
Acupuncture has demonstrated efficacy as a treatment for fibromyalgia; however, predictors of short- and long-term analgesic response in this population are not well understood.
Methods:
This manuscript describes a secondary analysis of a single-center, blinded, sham-controlled, randomized longitudinal acupuncture clinical trial in fibromyalgia. Baseline characterization included pressure pain threshold and pain interference, while residualized change in pain intensity from baseline to follow-up served as the primary outcome measure. Participants were randomized into groups that received verum (n = 36) or sham (n = 29) acupuncture treatment over a 12-week period (18 treatments) and were followed for 37 weeks from the initiation of treatment.
Results:
Lower pressure pain thresholds at baseline were associated with greater analgesia only in the sham treatment group immediately following treatment, while those with higher pressure pain thresholds had greater analgesia with verum treatment (B = −13.43, P = 0.001). Additionally, greater perceived impact of pain at baseline was predictive of greater short-term analgesia irrespective of treatment. Pressure pain threshold was not found to be predictive of long-term differential treatment response (B= −1.71, P = 0.66). There was a significant difference in the relationship between perceived impact of pain at baseline and subsequent long-term analgesia between groups where those with greater perceived impact of pain displayed improved long-term analgesia for verum acupuncture compared to the sham group (B = −11.37, P = 0.004).
Conclusions:
Our results support the use of a self-reported pain outcome in predicting long-term analgesia following acupuncture in fibromyalgia.
Keywords: Acupuncture, Fibromyalgia, Chronic pain
1. Introduction
Fibromyalgia (FM), the prototypical widespread nociplastic pain condition, is a prevalent and burdensome disease with few effective treatments [1]. FM is often associated with fatigue, sleep disturbance, disrupted cognition, and interference with daily living, in addition to chronic pain [1–4]. Acupuncture is increasingly used in FM [7], and some data supports that long-term pain reduction follows a course of treatment for several chronic pain conditions which overlap with FM (e.g., migraine, chronic low-back pain) [8,9]. While predictors of short-term response are beginning to be understood (see below), much less is known about individual factors that may predict long-term pain response to acupuncture. This would be of great clinical benefit, as some patients that undergo a course of treatment do not require additional sessions for months, and other patients require continued treatment to control symptoms. Little is known about how to predict these two outcomes prior to initiation of treatment.
In our prior studies we found that verum acupuncture was more effective in the short-term for individuals with higher pain thresholds, while sham acupuncture was more effective in individuals with lower pain thresholds [10,11]. In a related study [12], Baeumler et al. found that an initial clinical response to acupuncture (e.g., immediately following a single acupuncture treatment) was related to increased temporal summation of pain. However, these studies have not assessed factors that can predict whether a patient will have a sustained response to acupuncture treatment.
Studies that have explored patient demographics as predictors of the response to acupuncture treatment have not produced reliable findings: factors such as age and psychological distress are found to be significant in some studies, but no clinically meaningful differences have been determined in treatment response of pain to verum or sham acupuncture [8,13,14]. In a meta-analysis by Witt et al. [14], baseline psychological distress showed statistically significant but not clinically relevant moderation of chronic pain in response to acupuncture treatment, when compared to no treatment and sham acupuncture. Older patients trended toward showing better treatment response to acupuncture versus non-acupuncture treatment, but this effect was not statistically significant. In a single-session sham-controlled acupuncture trial, Wasan et al. [13] examined baseline psychiatric comorbidity and found no modification of response to verum or sham acupuncture treatment. Sherman et al. [8] also examined the characteristics of patients with chronic back pain who benefit from acupuncture versus sham treatment. They found that baseline quality of life measures (e.g., physical functioning, social functioning, pain, role emotional, mental health, etc.) were not significantly associated with differences in change in disability or symptom bothersomeness scores in either short- or long-term response, for up to 52 weeks. Kong et al. [15] found that the presence of coping strategies was associated with better treatment response of disability to acupuncture, but the same has not been found with respect to pain outcomes. Furthermore, even less is known about predictors of long-term response. Further examination of individual patient characteristics, particularly individual psychological factors as was focused on in the above studies, is needed to replicate prior findings and to characterize long-term response to acupuncture.
Here we investigated whether demographic, physical and psychological characteristics predict differential short-term and long-term pain reduction following acupuncture in FM in a secondary analysis of our prior larger randomized trial [16]. Among the patient-reported outcomes that we assessed at baseline, were those related to pain (e.g., intensity and interference) and other pain-related symptoms cardinal to FM (e.g., fatigue) that appear to improve with acupuncture treatment. We assessed which of these baseline characteristics predicted short-term and long-term pain reduction following verum acupuncture with sham treatment as a control. As we previously found that short-term response to acupuncture was associated with lower pain thresholds at baseline in this data [11], we also expected to replicate this finding in this distinct analysis.
2. Methods
2.1. Study protocol
The present study is a secondary analysis of a single center, blinded, sham-controlled, randomized longitudinal acupuncture clinical trial (NCT00010504) carried out at Georgetown University (Washington, DC USA) from August 2000 through January 2002. Study protocols were approved by the Georgetown University Institutional Review Board (IRB registration number: 0000193) and the study was conducted in accordance with the Declaration of Helsinki. All study participants provided written informed consent. This current study is a subsequent analysis from a previously published manuscript that examined acupuncture protocols; the original publication is henceforth referred to as the index study [16]. Short-term predictors of response to acupuncture in this study have also been analyzed and reported [11]. The analyses presented here, predicting long-term treatment response over 37 weeks, have yet to be reported.
2.2. Participants and procedure
Eligible and willing participants with FM based on the 1990 American College of Rheumatology criteria [17] for diagnosis of FM signed an informed consent document and were randomized into one of four treatment arms: traditional needle site with stimulation, traditional site without stimulation, non-traditional site with stimulation, and non-traditional needle site without stimulation. The original trial used a 2 × 2 factorial design to examine the individual and synergistic effects of needle location and stimulation. For complete details of this protocol see the index study [16]. For the purposes of this study, the traditional site with stimulation and traditional site without stimulation were combined and interpreted as verum acupuncture while the two non-traditional needle site sham groups (with and without stimulation) were combined as a single sham control group. Participants completed surveys at baseline prior to treatment, immediately after treatment completion (13 weeks), and at long-term follow-up (37 weeks). Our final sample of 65 participants had complete week 37 data. The surveys, detailed below and in the index study, included assessment of demographics, pain intensity, fatigue severity, physical health, catastrophizing, and expectations of acupressure. All predictor variables were assessed at the baseline assessment; pain intensity was assessed at all three timepoints.
2.3. Outcome measure
Pain intensity was assessed with a single-item 101-point numeric rating scale which ranged from 0 to 100 points in 5-point increments. The scale was anchored at 0 as “no pain” and 100 as “worst pain imaginable.”
2.4. Predictors
2.4.1. Pressure pain threshold
Participants underwent pressure pain testing at the non-dominant thumb at baseline prior to treatment to assess pain sensitivity. As described in prior analyses of this study [11], quantitative sensory testing was conducted using a multiple random staircase method to determine pain threshold [10]. Briefly, a pressure inducer was used to increase pressure on the participant’s thumbnail in discreet pressure and time intervals. Pain intensity ratings for each pressure were noted on a 21-point scale [18]. First, the participant’s pain sensitivity range was obtained by inducing stimuli of 5 s duration in increasing order with an interstimulus interval of 20 s. The initial stimulus was 0.25 kg/cm2 and increased by 0.25–0.50 kg/cm2 increments each time until either the participant’s maximum pain tolerance or 10 kg/cm2 was reached. Next, software titrated three independent series of increasing stimuli based on the method of multiple random staircases [19]. Each staircase was titrated to elicit pain sensations that were rated between 0 and 1 (faint pain), between 7 and 8 (mild to moderate pain), and between 13 and 14 (strong to slightly intense pain) on the 21-box numerical descriptor scale as previously described; sensations rated 15 and above were not elicited in this study. Each staircase delivered 12 five-second duration stimuli at 20 s intervals. Only stimulus intensities (in kg/cm2) obtained from the middle staircase (mild to moderate pain) were used. This middle staircase has shown to be more associated with clinical pain in this population [20]. Smaller values of pressure pain threshold indicate greater pressure pain sensitivity.
2.4.2. Demographic information
Participants self-reported their sex, race and age.
2.4.3. Perceived impact of pain
Perceived impact of pain was assessed with the four-item impact subscale of the Meaning of Illness Questionnaire [21]. The Meaning of Illness Questionnaire was developed as a means of assessing patients’ primary appraisal of their illness and the secondary appraisal of coping resources available to deal with the illness. The instrument possesses strong psychometric properties for assessing these constructs in research settings and has been empirically validated [22]. The impact subscale assessed the extent to which FM negatively affected daily life, relationships with family and friends, and work or school activities. Higher scores reflect a greater perceived impact of pain.
2.4.4. Fatigue severity
Fatigue was assessed with a single-item 0 (no fatigue) to 100 (worst fatigue imaginable) Visual Analogue Scale.
2.4.5. Physical health
Physical health was assessed with the Physical Components Summary score calculated from the Short Form-36 Health Survey [23]. This score ranges from 0 to 100 with higher scores indicating better health.
2.4.6. Catastrophizing
The 6-item catastrophizing subscale of the Coping Strategies Questionnaire [24] was used to assess the extent to which participants endorsed catastrophic thinking when experiencing symptoms on a 0 “never do that” to 6 “always do that” scale. Higher scores reflect greater catastrophizing.
2.4.7. Expectations of acupuncture
Expectancy was assessed with a single-item 7-point Likert scale measure that asked participants to rate “In general, what is your impression of acupuncture?” on a 1 “very negative” to 7 “very positive” scale.
2.5. Analysis
SPSS version 28.1.1 (Chicago, IL USA) was used to visualize and analyze data [25]. First, descriptive statistics were computed to characterize the study sample and examine assumptions necessary for analysis. Second, we calculated residualized change scores to create outcome measures for short- and long-term change in pain intensity as follows. Linear regression was used to predict pain intensity scores at 13 weeks from baseline scores (immediately after treatment, i.e., short-term change), and pain intensity at 37 weeks from baseline scores (six months after treatment, i.e., long-term change). Unstandardized residual values were saved as new variables and used as outcome variables in subsequent analyses. These residual values represent change in pain intensity, as they represent the variability in post-treatment pain intensity after accounting for variability predicted by baseline pain intensity [26]. Like raw change scores, unstandardized residualized change scores are in the same metric as the original measure. Negative values represent reduction in pain intensity, zero represents no change, and positive values represent increase in pain. Descriptive statistics were also computed and examined for violation of assumptions.
Between-group comparisons (sham versus verum) were conducted to assess whether groups had differences in demographics or potential moderating variables at baseline. Chi square tests and t-tests were performed on categorical and continuous variables, respectively. Groups with the largest proportion of cases were selected as reference groups. Race categories were collapsed into White and non-White as there were too few cases in the non-White categories to conduct comparisons.
For the primary analyses, ordinary least squares regression with listwise deletion was performed with the SPSS regression procedure. Moderation analyses were used to test whether the baseline predictors individually influenced the effect of treatment group on change in pain intensity. To facilitate interpretation of estimates, continuous predictor variables were centered at the grand mean. Backward stepwise variable selection was used for our model-building procedure. This procedure occurred in two stages. In the first stage, a backward stepwise regression procedure was used on an initial model limited only to demographic variables [sex (0 = female, 1 = male), race (0 = White; 1 = non-White), and age] predicting residualized change in pain intensity. The model was trimmed one term at time, with the term with largest P value removed at each step until only terms with P < 0.10 were retained. In the second stage, the backward stepwise procedure was used on an expanded model that contained demographic variables retained from the prior stage, treatment group (0 = sham; 1 = verum), main effect terms for all potential moderators (mild to moderate pressure pain threshold, perceived impact of pain, fatigue severity, physical health, catastrophizing, and acupuncture expectations), and each potential moderator’s interaction with treatment group. Backward elimination was performed on this model with some caveats. Treatment group was retained in all steps, and main effect terms of the potential treatment moderators were retained if their corresponding interaction with treatment was still in the model at a given step, regardless of P value. Statistically significant (P < 0.05) moderator by interaction terms was interpreted as an indication of differential treatment effects, that is, that change in pain differed between study groups as a function of the predictor of interest. Significant interactions in the final model were probed with simple slopes at ± 1 standard deviation. R2 change estimates were computed to obtain measures of effect size. R2 values of 0.01, 0.09 and 0.25 were interpreted as small, medium and large effect sizes, respectively [27]. This procedure was conducted separately for each outcome: residualized change in pain intensity at post-treatment and long-term (6 months) follow-up.
3. Results
3.1. Sample characteristics and between-group comparisons
A description of the overall sample can be found in the index study [16]. In summary, participants tended to be middle-aged, female, and of White/Caucasian descent (N = 65). Twenty-nine participants received sham acupuncture treatment and 36 received verum acupuncture. All assumptions required for data analysis were met for all study variables.
Table 1 presents sample characteristics and descriptive statistics by treatment group and a summary of bivariate comparisons between the sham and verum treatment groups. Participants in the two treatment groups did not appear to differ on any study measures at baseline.
Table 1.
Descriptive statistics by treatment group and between-group comparisons
| Variable | Possible min and max | Sham acupuncture (n = 29) | Verum acupuncture (n = 36) | P value | ||
|---|---|---|---|---|---|---|
|
| ||||||
| n | Value | n | Value | |||
|
| ||||||
| Demographics | ||||||
| Sex (n [%]) | / | 0.10 | ||||
| Female (reference) | / | 25 | 86.2 | 35 | 97.2 | |
| Male | / | 4 | 13.8 | 1 | 2.8 | |
| RaceA (n [%]) | / | 1.00 | ||||
| White (reference) | / | 26 | 89.7 | 33 | 91.7 | |
| Black | / | 3 | 10.3 | 2 | 5.6 | |
| Asian | / | 0 | 0.0 | 1 | 2.8 | |
| Age (mean ± SD, years) | 26, 64 | 29 | 49.3 ± 9.5 | 36 | 47.5 ± 9.6 | 0.45 |
| Age (n [%]) | ||||||
| 24–34 | / | 3 | 10.3 | 4 | 11.1 | |
| 35–44 | / | 4 | 13.8 | 7 | 19.4 | |
| 45–54 | / | 12 | 41.4 | 17 | 47.2 | |
| 55–64 | / | 10 | 34.5 | 8 | 22.2 | |
| Potential treatment moderators (mean ± SD) | ||||||
| Baseline mild to moderate pressure pain threshold* (kg/cm2) | 0, 10 | 29 | 2.39 ± 1.62 | 36 | 2.73 ± 1.76 | 0.42 |
| Baseline impact of pain (MIQ Impact) | 0, 6 | 28 | 2.15 ± 1.62 | 35 | 2.03 ± 1.68 | 0.77 |
| Baseline fatigue severity (VAS) | 0, 100 | 29 | 61.21 ± 25.02 | 36 | 49.72 ± 24.38 | 0.07 |
| Baseline physical health (SF-36 PCS) | 0, 100 | 28 | 38.36 ± 6.88 | 35 | 36.13 ± 8.41 | 0.26 |
| Baseline catastrophizing (CSQ) | 0, 36 | 28 | 2.05 ± 1.22 | 36 | 1.63 ± 0.90 | 0.12 |
| Baseline acupuncture expectations (higher = more positive) | 1, 7 | 27 | 5.15 ± 1.20 | 35 | 5.49 ± 1.09 | 0.25 |
| Pain intensity measures (mean ± SD) | ||||||
| Pain intensity at baseline | 0, 100 | 29 | 55.86 ± 23.61 | 36 | 52.78 ± 22.72 | 0.59 |
| Pain intensity at post-treatment | 0, 100 | 27 | 47.59 ± 25.62 | 33 | 40.45 ± 30.22 | |
| Pain intensity at long-term follow-up | 0, 100 | 24 | 48.33 ± 23.67 | 30 | 54.67 ± 28.80 | |
| Residualized change in pain intensity at post-treatmentB | -100, 100 | 27 | 2.98 ± 24.57 | 33 | -2.44 ± 30.02 | |
| Residualized change in pain intensity at long-term follow-upB | -100, 100 | 24 | -3.83 ± 22.61 | 30 | 3.06 ± 27.90 | |
Assessments occurred at week 0 (baseline), week 13 (post-treatment), and week 37 (long-term 6-month follow-up); higher = more of construct unless otherwise indicated.
non-White groups collapsed into 1 group.
lower pressure pain threshold scores associated with more pain sensitivity.
baseline scores were regressed on post-treatment and longterm follow-up scores, respectively; resulting residualized scores were saved and used in subsequent analyses; lower = greater reduction in pain. CSQ: Coping Strategies Questionnaire; MIQ: Meaning of Illness Questionnaire; SD: standard deviation; SF-36 PCS: Short Form Health Survey-36 Physical Component Summary; VAS: Visual Analogue Scale.
Considering all possible data points (12 study variables × 65 participants), data in the sample were 96.9% complete (756 data points/780 possible data points). Forty-six participants had complete data, 17 participants were missing 1 data point, 1 participant was missing 3 data points, and 1 participant was missing 4 data points. No participants were missing the week 37 clinical pain outcome. Participants excluded from the short-term analgesic response analysis due to missing data (n = 7) did not differ statistically from participants included in the analysis (n = 58) demographically or with respect to the other study measures. Similarly, participants excluded from the long-term analgesic response analysis (n = 13) did not differ statistically from those who were included (n = 52) on the same measures.
3.2. Moderation of treatment effect
3.2.1. Short-term analgesic response
In the initial backward stepwise procedure, no demographic variables were significantly associated with reduction in pain intensity after treatment and were thus excluded from subsequent iterations of the model (age removed at P = 0.72, sex at P = 0.49, and race at P = 0.18). After expanding the model to include all possible moderators in the second round of backward elimination, a number of variables were removed from the final model (see Table 2 for additional information), resulting in the model presented in Table 3.
Table 2.
Terms eliminated from expanded model through backward stepwise procedure.
| Step | Short-termA analgesic response model | Long-termB analgesic response model | ||
|---|---|---|---|---|
|
| ||||
| Term eliminated | P value | Term eliminated | P value | |
|
| ||||
| 1 | NoneC | N/A | NoneC | N/A |
| 2 | Treatment group by fatigue severity interaction | 0.59 | Treatment group by catastrophizing interaction | 0.99 |
| 3 | Treatment group by acupuncture expectation interaction | 0.47 | Treatment group by acupuncture expectation interaction | 0.66 |
| 4 | Acupuncture expectation main effect | 0.60 | Treatment group by fatigue severity interaction | 0.67 |
| 5 | Treatment group by catastrophizing interaction | 0.36 | Treatment group by mild to moderate pressure pain threshold interaction | 0.66 |
| 6 | Catastrophizing main effect | 0.93 | Mild to moderate pressure pain threshold main effect | 0.84 |
| 7 | Fatigue severity main effect | 0.21 | Fatigue severity main effect | 0.49 |
| 8 | Treatment group by physical health interaction | 0.16 | Acupuncture expectation main effect | 0.30 |
| 9 | Physical health main effect | 0.30 | Catastrophizing main effect | 0.26 |
| 10 | Treatment group by impact of pain interaction | 0.27 | Treatment group by physical health interaction | 0.17 |
immediately post-treatment, 13 weeks after study initiation.
at long-term 6-month follow-up, 37 weeks after study initiation.
full model, no terms eliminated. N/A: not applicable. Treatment group: sham vs verum acupuncture.
Table 3.
Parameter estimates for final regression models predicting short- and long-term analgesic response to sham versus verum acupuncture
| Item | B | SE | P value | LL CI | UL CI | R 2 |
|---|---|---|---|---|---|---|
|
| ||||||
| A. Short-term analgesic response model | ||||||
| Intercept | 4.74 | 4.70 | 0.32 | −4.68 | 14.16 | |
| Treatment group | −8.53 | 6.31 | 0.18 | −21.19 | 4.13 | 0.024 |
| Baseline impact of pain | −6.37 | 1.96 | 0.002 | −10.30 | −2.43 | 0.138 |
| Baseline mild to moderate pressure pain threshold | 5.25 | 2.98 | 0.08 | −0.74 | 11.23 | 0.041 |
| Baseline mild to moderate pressure pain threshold*treatment group | −13.43 | 3.88 | 0.001 | −21.21 | −5.66 | 0.157 |
| B. Long-term analgesic response model | ||||||
| Intercept | −3.01 | 4.59 | 0.52 | −12.24 | 6.23 | |
| Treatment group | 4.32 | 6.21 | 0.49 | −8.17 | 16.81 | 0.007 |
| Baseline physical health | −0.79 | 0.45 | 0.08 | −1.69 | 0.11 | 0.045 |
| Baseline impact of pain | 3.05 | 2.98 | 0.31 | −2.94 | 9.04 | 0.015 |
| Baseline impact of pain*treatment group | −11.37 | 3.72 | 0.004 | −18.86 | −3.88 | 0.134 |
The outcome was residualized change in pain intensity at post-treatment (13 weeks after study initiation) and at a long-term 6-month follow-up (37 weeks after study initiation), relative to baseline; possible range −100 to 100; lower = greater reduction in pain. Treatment group: 0 = sham; 1 = verum. Measures (minimum and maximum possible values): impact of pain, Meaning of Illness Questionnaire Impact subscale (0, 6); physical health, SF-36 Physical Component Summary (0, 100); mild to moderate pressure pain threshold, kg/cm2 (0, 10) with lower values mean higher pain sensitivity. Higher values = more of construct unless otherwise specified. LL CI: lower limit of confidence interval; SE: standard error; UL CI: upper limit of confidence interval.
A significant pressure pain threshold-by-treatment group interaction effect was identified: among participants who received sham treatment, those demonstrating lower pressure pain thresholds at baseline showed a trend towards greater reduction in pain intensity at post-treatment relative to those with higher pressure pain thresholds. Among participants who received the verum treatment, those demonstrating higher pressure pain thresholds reported greater reduction in pain intensity compared to those demonstrating lower pressure pain thresholds. See Fig. 1 for scatter plots of baseline pressure pain thresholds and subsequent change in clinical pain for verum and sham groups.
Fig. 1.
Change in pain at week 13 for verum (A) and sham (B) acupuncture group. Higher baseline mild to moderate pressure pain thresholds were associated with greater short-term analgesia in the verum acupuncture group, while lower pressure pain thresholds were associated with short-term analgesia in the sham acupuncture group, though the latter result did not reach statistical significance (P < 0.05). Outcome: residualized change in pain intensity visual analogue rating (possible minimum, maximum: −100, 100), lower = greater reduction in pain.
Additionally, a significant main effect for perceived impact of pain was found. Participants who reported greater perceived impact of pain at baseline were more likely to report greater reduction in pain, regardless of treatment group. The r2 values of both effects corresponded with medium effect sizes.
3.2.2. Long-term analgesic response
No demographic variables were associated with reduction in pain intensity at the long-term follow-up and were consequently removed from the model via backward elimination in the initial round of backward stepwise regression (race removed at P = 0.87, sex at P = 0.50, and age at P = 0.21). In the second round of backward elimination, several variables were removed from the expanded model (see Table 2), resulting in the model shown in Table 3.
A significant impact of pain-by-treatment group interaction was found. In the verum group, participants who reported a greater perceived impact of pain at baseline were more likely to report higher reduction in pain intensity at the long-term follow-up relative to participants who initially reported a lesser perceived impact of pain. In contrast, perceived impact of pain at baseline did not appear to be associated with change in pain intensity in the sham group. See Fig. 2 for scatter plots of baseline perceived impact of pain and subsequent change in clinical pain for verum and sham groups. Though it was not significant at P < 0.05, a main effect trend (P < 0.10) correlated to baseline physical health was also present: participants who reported better physical health at baseline were marginally more likely to report a higher reduction in pain intensity than those reporting worse baseline physical health, regardless of treatment group. The r2 values of the interaction and main effects trend correspond with medium and small-to-medium effect sizes, respectively.
Fig. 2.
Change in pain at week 37. Higher impact of pain was associated with greater long-term analgesia in the verum acupuncture treatment group, whereas no association was found between impact of pain and change in pain intensity in the sham acupuncture group. Outcome: residualized change in pain intensity visual analogue rating (possible minimum, maximum: −100, 100), lower means greater reduction in pain. n.s. means not significant.
4. Discussion
The purpose of this study was to identify predictors of short- and long-term analgesic response to verum versus sham acupuncture. In the short-term, those with increased pressure pain sensitivity (i.e., lower pain thresholds) at baseline had greater analgesia in the sham treatment group, while those with lower pressure pain sensitivity (i.e., higher pain thresholds) had greater analgesia with verum treatment. Pressure pain thresholds were not found to be predictive of long-term outcomes. In the long term, those with greater perceived impact of pain had improved analgesia in the verum group, while in the sham group there was no relationship between perceived impact of pain at baseline and subsequent change in clinical pain. Interestingly, in the short term, those with greater perceived impact of pain had greater analgesia in both verum and sham treatment groups.
Our finding that baseline pressure pain threshold was a differential predictor of short-term analgesia following verum and sham acupuncture treatment is generally consistent with the results of multiple prior studies. For example, a 2013 FM study by Harte et al. [10] showed that individuals who received sham acupuncture were more likely to benefit after treatment if they demonstrated lower pressure pain thresholds at baseline compared to those with higher pain thresholds. Similarly, in 2017, secondary analysis of the index study by Zucker et al. [11] showed that individuals with FM with higher pressure pain thresholds had more short-term analgesia with verum acupuncture treatment and individuals with lower pressure pain thresholds had more short-term analgesia with sham acupuncture. Other studies have shown that the initial clinical response to acupuncture is related to increased temporal summation of pain at baseline [12]. Though we were able to show that pressure pain thresholds correlated with treatment response in the short-term in the current study, we did not find evidence that this effect persisted at the long-term follow-up.
In contrast, we found that the perceived impact of pain on daily life prior to treatment was a differential predictor of long-term analgesia following verum or sham acupuncture. This finding, coupled with the result that greater impact of pain was associated with short-term analgesia irrespective of treatment, suggests that initial perceptions of functional limitation are deeply important for treatment success, non-specifically in the short term and specifically for verum acupuncture in the long term. This is a potential new finding for the field and may be important in clinical practice. Perceived impact of pain, a subscore from the Meaning of Illness Questionnaire, is a part of the cognitive appraisal of a stressful event, in this case FM pain. In the past, Meaning of Illness Questionnaire has been validated to explain adjustment to illness with better adjustment relating to lesser perceived impact of pain [28]. In contrast, those with greater perceived pain impact may display greater pain interference with life events. This interference of pain on functioning (e.g., family relations, work, etc.) may also be representative of a broader level of psychological distress based on multiple factors. It might be possible that having more functional limitations makes individuals more prone to respond to verum acupuncture. These individuals may have an increased need for treatment relief and a subsequent enhanced response to acupuncture.
Neurobiological differences between short- and long-term changes in pain may also contribute to these differences. One such mechanism is synaptic plasticity which is modulated during acupuncture analgesia. This mechanism may explain the interaction of central pain thresholds and temporal analgesic outcomes following acupuncture [29–31]. For example, reductions in synaptic strength, as seen with increases in long-term depression, may reduce neuronal coupling between pro-nociceptive neurons [32]. Thus, in individuals with increased synaptic plasticity (i.e. increased long-term depression), one could expect improved analgesia response following acupuncture via reduction in nociceptive pain transmission by reducing coupling between pain-promoting neurons. Long-term benefits of acupuncture are also likely to have impacts on gene expression and proteomics, and if individuals have differing gene pathways, this may explain differences in temporal response to acupuncture treatment [33–35]. For example, animal models demonstrate the impacts of acupuncture on specific microRNA, which impacts gene expression and subsequent synaptic plasticity [36]. Additional work on the examination of synaptic plasticity in depression has demonstrated the impact of acupuncture on neurotransmitter receptors (i.e., serotonin) in animal models [37]. While MacPherson et al. [9] showed that pain reduction can persist for months following acupuncture treatment, the underlying neurobiological determinants are still being investigated.
The findings in this study have clinical implications that can be applied to how acupuncture is practiced. Assessing pain pressure thresholds may be helpful in identifying which specific patients will initially respond to verum acupuncture treatment. Our results suggest that patients with higher pain thresholds may respond well to verum acupuncture, while for those reporting conditions like frequent/severe allodynia, verum treatments may be contraindicated. Though pain pressure thresholds did not appear to have a differential impact on long-term analgesia, it may be possible that low pain thresholds could potentially adversely affect treatment continuation in the short-term. Assessing the extent to which pain disrupts daily life with a perceived impact of pain questionnaire may help determine which specific patients may benefit in the long term from this therapy. This type of information would be of great value to know prior to embarking on a course of therapy that may involve months of treatments. Better treatment planning could help divert resources to those who would benefit most.
While the findings in this work are significant and robust, there are a few limitations that should be acknowledged. First, this sample size was relatively small. This means that we did not have statistical power to detect small changes. Second, our sample was also not representative of the general population. We had a predominately white and female population in our study. In future studies, given the complexity of the construct of fatigue, a more robust scale to assess fatigue would be additive. Our study is also a secondary analysis of a previous trial, and the findings presented here were not the original primary aims of this study. Finally, other possible predictors may exist that we did not capture in this analysis.
5. Conclusion
In summary, this study replicated prior findings showing pain pressure thresholds to be a predictor of short-term but not long-term analgesia. We also find that FM participants with greater baseline perceived impact of pain report better long-term analgesia at 6 months following verum acupuncture treatment. Taken together, these results suggest that predictors of short- and long-term acupuncture treatment response should not be assumed to be the same. In the future, further study of long-term response to acupuncture treatment in FM is needed, and factors which impact long-term response need further elucidation. It remains to be seen if our findings can be generalized to other chronic pain conditions.
Funding
This project was supported by National Institutes of Health (NCCAM) grant R01 AT00004, Department of Defense DAMD-17002–0018, and Georgetown University GCRC grant 98–186. REH was supported by National Institutes of Health (NCCAM) K01 AT01111–01 and National Institutes of Health grant R01-AT-00755.
Footnotes
Declaration of competing interest
The authors have no conflicts of interest to declare.
CRediT authorship contribution statement
AEW, AA, REH and DAW designed the hypotheses and the experiments. AA and REH performed the experiments and their analyses. REH was responsible for data collection. AEW, AA and REH participated in data analysis. All authors participated in data interpretation and manuscript review and writing. AEW, AA, LTS and REH were responsible for preparation of data tables and figures. All authors contributed to the scientific discussion of the data and of the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Clauw DJ. Fibromyalgia: a clinical review. JAMA 2014;311(15):1547–55. [DOI] [PubMed] [Google Scholar]
- 2.Bennett RM, Jones J, Turk DC, Russell IJ, Matallana L. An internet survey of 2,596 people with fibromyalgia. BMC Musculoskelet Disord 2007;8:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.McBeth J, Jones K. Epidemiology of chronic musculoskeletal pain. Best Pract Res Clin Rheumatol 2007;21(3):403–25. [DOI] [PubMed] [Google Scholar]
- 4.Skaer TL. Fibromyalgia: disease synopsis, medication cost effectiveness and economic burden. Pharmacoeconomics 2014;32(5):457–66. [DOI] [PubMed] [Google Scholar]
- 5.Vincent A, Lahr BD, Wolfe F, Clauw DJ, Whipple MO, Oh TH, et al. Prevalence of fibromyalgia: a population-based study in Olmsted County, Minnesota, utilizing the Rochester Epidemiology Project. Arthritis Care Res (Hoboken) 2013;65(5):786–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wolfe F, Ross K, Anderson J, Russell IJ, Hebert L. The prevalence and characteristics of fibromyalgia in the general population. Arthritis Rheum 1995;38(1):19–28. [DOI] [PubMed] [Google Scholar]
- 7.Clarke TC, Black LI, Stussman BJ, Barnes PM, Nahin RL. Trends in the use of complementary health approaches among adults: United States, 2002—2012. Natl Health Stat Report 2015(79):1–16. [PMC free article] [PubMed] [Google Scholar]
- 8.Sherman KJ, Cherkin DC, Ichikawa L, Avins AL, Barlow WE, Khalsa PS, et al. Characteristics of patients with chronic back pain who benefit from acupuncture. BMC Musculoskelet Disord 2009;10:114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.MacPherson H, Vertosick EA, Foster NE, Lewith G, Linde K, Sherman KJ, et al. The persistence of the effects of acupuncture after a course of treatment: a meta-analysis of patients with chronic pain. Pain 2017;158(5):784–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Harte SE, Clauw DJ, Napadow V, Harris RE. Pressure pain sensitivity and insular combined glutamate and glutamine (Glx) are associated with subsequent clinical response to sham but not traditional acupuncture in patients who have chronic pain. Med Acupunct 2013;25(2):154–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zucker NA, Tsodikov A, Mist SD, Cina S, Napadow V, Harris RE. Evoked pressure pain sensitivity is associated with differential analgesic response to verum and sham acupuncture in fibromyalgia. Pain Med 2017;18(8):1582–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Baeumler PI, Conzen P, Irnich D. High temporal summation of pain predicts immediate analgesic effect of acupuncture in chronic pain patients-a prospective cohort study. Front Neurosci 2019;13:498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wasan AD, Kong J, Pham LD, Kaptchuk TJ, Edwards R, Gollub RL. The impact of placebo, psychopathology, and expectations on the response to acupuncture needling in patients with chronic low back pain. J Pain 2010;11(6):555–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Witt CM, Vertosick EA, Foster NE, Lewith G, Linde K, MacPherson H, et al. The effect of patient characteristics on acupuncture treatment outcomes: an individual patient data meta-analysis of 20,827 chronic pain patients in randomized controlled trials. Clin J Pain 2019;35(5):428–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kong JT, Puetz C, Tian L, Haynes I, Lee E, Stafford RS, et al. Effect of electroacupuncture vs sham treatment on change in pain severity among adults with chronic low back pain: a randomized clinical trial. JAMA Netw Open 2020;3(10):e2022787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Harris RE, Tian X, Williams DA, Tian TX, Cupps TR, Petzke F, et al. Treatment of fibromyalgia with formula acupuncture: investigation of needle placement, needle stimulation, and treatment frequency. J Altern Complement Med 2005;11(4):663–71. [DOI] [PubMed] [Google Scholar]
- 17.Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia. report of the Multicenter Criteria Committee. Arthritis Rheum 1990;33(2):160–72. [DOI] [PubMed] [Google Scholar]
- 18.Petzke F, Harris RE, Williams DA, Clauw DJ, Gracely RH. Differences in unpleasantness induced by experimental pressure pain between patients with fibromyalgia and healthy controls. Eur J Pain 2005;9(3):325–35. [DOI] [PubMed] [Google Scholar]
- 19.Gracely RH, Lota L, Walter DJ, Dubner R. A multiple random staircase method of psychophysical pain assessment. Pain 1988;32(1):55–63. [DOI] [PubMed] [Google Scholar]
- 20.Harris RE, Gracely RH, McLean SA, Williams DA, Giesecke T, Petzke F, et al. Comparison of clinical and evoked pain measures in fibromyalgia. J Pain 2006;7(7):521–7. [DOI] [PubMed] [Google Scholar]
- 21.McAdams C BC, Browne G. Meaning of Illness Questionnaire: questionnaire and scoring manual. Hamilton: McMaster University School of Nursing; 1989. [Google Scholar]
- 22.Browne GB, Byrne C, Roberts J, Streiner D, Fitch M, Corey P, et al. The meaning of illness questionnaire: reliability and validity. Nurs Res 1988;37(6):368–73. [PubMed] [Google Scholar]
- 23.Ware JE Jr, Gandek B. Overview of the SF-36 Health Survey and the International Quality of Life Assessment (IQOLA) Project. J Clin Epidemiol 1998;51(11):903–12. [DOI] [PubMed] [Google Scholar]
- 24.Hirsh AT, George SZ, Riley JL, Robinson ME. An evaluation of the measurement of pain catastrophizing by the coping strategies questionnaire. Eur J Pain 2007;11(1):75–81. [DOI] [PubMed] [Google Scholar]
- 25.Corp I. IBM SPSS Statistics. Chicago: IMB Corp; 2021. [Google Scholar]
- 26.MacKinnon DP. Introduction to statisical mediation analysis. New York: Taylor & Francis Group; 2008. [Google Scholar]
- 27.Cohen J Statistical power analysis for the behavioral sciences. 2nd ed. New York: Routledge; 1988. [Google Scholar]
- 28.Weir R, Browne G, Roberts J, Tunks E, Gafni A. The Meaning of Illness Questionnaire: further evidence for its reliability and validity. Pain 1994;58(3):377–86. [DOI] [PubMed] [Google Scholar]
- 29.Kan B, Dong Z, Tang Z, Zhao L, Li Z. Acupuncture improves synaptic plasticity of SAMP8 mice through the RhoA/ROCK pathway. Curr Alzheimer Res 2023;20(6):420–30. [DOI] [PubMed] [Google Scholar]
- 30.Lyu Z, Guo Y, Gong Y, Fan W, Dou B, Li N, et al. The role of neuroglial crosstalk and synaptic plasticity-mediated central sensitization in acupuncture analgesia. Neural Plast 2021;2021:8881557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Yang P, Chen HY, Zhang X, Wang T, Li L, Su H, et al. Electroacupuncture attenuates neuropathic pain in a rat model of cervical spondylotic radiculopathy: involvement of spinal cord synaptic plasticity. J Pain Res 2023;16:2447–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Xing GG, Liu FY, Qu XX, Han JS, Wan Y. Long-term synaptic plasticity in the spinal dorsal horn and its modulation by electroacupuncture in rats with neuropathic pain. Exp Neurol 2007;208(2):323–32. [DOI] [PubMed] [Google Scholar]
- 33.Yin LM, Wang Y, Wang Y, Xu YD, Liu YY, Jin WR, et al. Effects of acupuncture on the gene expression profile of lung tissue from normal rats. Mol Med Rep 2012;6(2):345–60. [DOI] [PubMed] [Google Scholar]
- 34.Xu YD, Wang Y, Park GH, Yin LM, Ran J, Liu YY, et al. Non-specific physiological background effects of acupuncture revealed by proteomic analysis in normal rats. BMC Complement Altern Med 2014;14:375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Oh JE, Kim SN. Anti-inflammatory effects of acupuncture at ST36 point: a literature review in animal studies. Front Immunol 2021;12:813748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Dai Y, Wang S, Yang M, Zhuo P, Ding Y, Li X, et al. Electroacupuncture protective effects after cerebral ischemia are mediated through miR-219a inhibition. Biol Res 2023;56(1):36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Han X, Wu H, Yin P, Chen Z, Cao X, Duan Y, et al. Electroacupuncture restores hippocampal synaptic plasticity via modulation of 5-HT receptors in a rat model of depression. Brain Res Bull 2018;139:256–62. [DOI] [PubMed] [Google Scholar]