Efficacy and Safety of Auricular Acupuncture for the Treatment of Insomnia in Breast Cancer Survivors: A Randomized Controlled Trial

Simple Summary Sleep problems impair up to 70% of breast cancer survivors. We studied whether ear acupuncture improves sleep quality in breast cancer survivors with sleeplessness. Fifty-two female breast cancer survivors were randomly distributed to either 10 group ear acupuncture sessions or a onetime psychoeducation group. We found that sleep quality improved over the 5 weeks of the study in the ear acupuncture group in comparison to the psychoeducation group. Stress, anxiety and fatigue improved as well. These effects were lost after 17 and 29 weeks. Abstract Among women, breast cancer is the most commonly diagnosed cancer worldwide. Sleep problems impair 40–70% of breast cancer survivors. This randomized controlled trial evaluates the effect of auricular acupuncture on sleep quality in breast cancer survivors suffering from insomnia. Fifty-two female breast cancer survivors with insomnia (mean age 55.73 ± 8.10 years) were randomized either to 10 treatments of auricular acupuncture within five weeks (n = 26), or to a single session of psychoeducation plus an insomnia advice booklet (n = 26). The primary outcome was sleep quality (measured by the Pittsburgh Sleep Quality Index) at week 5. Secondary outcomes were inflammation parameter (interleukin-6), stress, anxiety, depression, quality of life, and fatigue at week 5, and sleep quality, stress, anxiety, depression, quality of life, and fatigue 17 and 29 weeks after randomization. Intention-to-treat analysis showed a significantly stronger increase in sleep quality in the auricular acupuncture group compared to the psychoeducation group (p = 0.031; η2p = 0.094) at week 5. Furthermore, auricular acupuncture improved stress (p = 0.030; η2p = 0.094), anxiety (p = 0.001; η2p = 0.192), and fatigue (p = 0.006; η2p = 0.148) at week 5 compared to psychoeducation. No significant group difference was found concerning the other outcomes at week 5, or in any outcome at week 17 or week 29. No serious adverse events occurred during the study period. In conclusion, a semi-standardized group auricular acupuncture might be an effective and safe intervention in treating insomnia in breast cancer survivors in the short term, and may reduce stress, anxiety, and fatigue as well. Long-term effects remain questionable.


Introduction
Among females, breast cancer is the most commonly diagnosed cancer, with an incidence of 11.6% worldwide in 2018 [1], and a 10-year survival rate exceeding 80% [2]. About 40-70% of breast cancer survivors report sleep problems even years after diagnosis, was pre-screened in a telephone interview. The final inclusion of participants, fulfilling all inclusion and none of the exclusion criteria (see Table 1), was performed by the study physician. Before enrollment, participants received detailed written and verbal information about the study, and signed a consent form indicating their agreement to participate. The participants did not receive financial compensation for participation. The conduction of the study was in adherence to the tenets of the Declaration of Helsinki [39] and to good clinical practice, and the report was written in accordance with the CONSORT 2010 Statement [40] and the CONSORT harm extension [41].

Randomization
Patients were randomized to auricular acupuncture or psychoeducation after the first data collection by means of block randomization with randomly varying block lengths. The randomization list was generated by HC using a computerized randomization generator [42]. The list was password secured, and neither the study physicians nor the data collectors had access to it. Allocation concealment was achieved by central randomization by a study nurse so that investigators enrolling participants could not foresee allocation, and further treatment was carried out according to group allocation. Neither the patient nor the investigator had any influence on the randomization.

Interventions
Two licensed acupuncturists (Petra Voiss and Wiebke Kohl), with 20 and 5 years of experience, respectively, used the "very point" method to insert 0.20 × 15-mm TeWa PB-Type needles. The patients were asked on which side they would like to receive auricular acupuncture. After skin disinfection and inspection of the ear, the acupuncturist worked with a semi-standardized protocol and placed the needles in sensitive points-first on the postantitragal belt, second on the helix channel, and finally on the shen men. Additional points were used to address comorbid symptoms such as anxiety, hot flushes, and pain ( Figure 1, Supplementary Table S1). The acupuncture needles were inserted 2-3 mm deep. Five to eight needles were used, and only one ear per patient was treated twice weekly for five weeks in a group setting of 6-10 participants. Patients were placed in reclining chairs in a group room at the hospital. The patients were asked to relax during the auricular acupuncture and keep quiet. The needles remained for at least 20 min. Comorbid conditions and adverse events were noted after each acupuncture session.
The control group received a single 90-min psychoeducation group session (8 patients) held by an experienced psychologist trained in psychoeducation/cognitive therapy for insomnia patients (Melanie D. Höxtermann). The program was based on the German S3 guidelines "Restless Sleep/Sleep Disorders" [43], and conveyed information concerning normal sleep, sleep problems, the use of sleep restriction in treating sleep problems, the use of stimulus control, psychohygiene, and cognitive behavioral elements that could be helpful in resolving sleep problems. Furthermore, all participants receiving psychoeducation received an advice booklet for dealing with sleep disorders [44].

Measures
All outcomes were assessed at weeks 0, 5, 17, and 29, except for IL-6-which was only assessed at weeks 0 and 5-and adverse events-which were continuously assessed during the study, and at weeks 5 and 17. Sociodemographic and clinical characteristics, as well as treatment expectations, were assessed at baseline.

Primary Outcome
Sleep Quality at Week 5 The primary outcome of this study was subjective sleep quality, measured by the Pittsburgh Sleep Quality Index (PSQI) [45] at the end of the intervention at week 5. The PSQI contains 19 self-assessment questions and 5 questions rated by the bed partner or roommate. The 5 questions rated by the bed partner or roommate are used as clinical information, and not included in quantitative analysis. The PSQI evaluates sleep quality in 7 domains over the past month (subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction). The total score is generated by the summation of the component scores, and can range from 0 to 21, with a higher score corresponding to a reduced sleep quality [45]. A cutoff score of 5 or 8 represents a "poor" sleeper in the general population or in cancer patients, respectively [46]. A change of 3 points or more on the PSQI is considered a minimal clinically important difference [47]. Brackhaus et al. showed the PSQI to have a high test-retest reliability and a good validity for patients with primary insomnia [48].
2.6. Secondary Outcomes 2.6.1. Sleep Quality For secondary outcomes, the PSQI was surveyed at week 17 and week 29.
2.6.2. IL-6 Plasma Levels IL-6 plasma levels were determined with the Human Th1/Th2 Cytokine Kit II (BD Biosciences, Heidelberg, Germany) at weeks 0 and 5. Capture beads were mixed, centrifuged for 5 min at 200× g, and the supernatant was aspirated and discarded. Bead pellets were vigorously resuspended in serum-enhancement buffer. For each sample-control and standard-50 µL of the capture bead suspension was added to a BD Falcon FACS tube (BD Biosciences, Heidelberg, Germany), followed by 50 µL of each standard, control, or ethylenediaminetetraacetic acid (EDTA) plasma sample and, subsequently, 50 µL of detection reagent. All tubes were incubated for 3 h at room temperature in the dark. After addition of 1 mL of wash buffer, samples were centrifuged for 5 min at 200× g. The supernatant was removed by aspiration, the bead pellets were resuspended in 300 µL of wash buffer, and the samples were acquired on a BD FACS Canto II with the FACS DIVA software 6.1.3 (both BD Biosciences, Heidelberg, Germany). Data were analyzed using the FCAP Array 3.0.19.2091 software (BD Biosciences, Heidelberg, Germany).

Anxiety and Depressive Symptoms
The Hospital Anxiety and Depression Scale (HADS) is a self-report questionnaire consisting of 14 items that measure psychological wellbeing on the two scales depression HADS-D (7 items) and anxiety HADS-A (7 items), each with a sum score range of 0-21 [49].
Higher scores indicate greater psychological burden. The HADS has been shown to be a valid instrument for detecting clinically meaningful results as a psychological screening tool, and is sensitive to changes in response to therapies [50]. Recent research suggests an optimal cutoff of >9 for the HADS-A and of >7 for the HADS-D in cancer patients [51]. The HADS was assessed at weeks 0, 5, 17, and 29.

Quality of Life and Fatigue
The Functional Assessment Of Cancer Therapy-Breast Cancer (FACT-B) [52] and Functional Assessment Of Chronic Illness Therapy-Fatigue (FACIT-F) [53] were secondary outcomes, and were assessed at weeks 0, 5, 17, and 29. The FACT-B uses the 27-item Functional Assessment of Cancer Therapy-General (FACT-G) questionnaire as a base, supplemented by 9 breast cancer items measured on a five-point Likert scale ranging from 0 ("not at all") to 4 ("very much so"). Higher scores indicate a better quality of life [52,54,55]. The FACIT-F subscale consists of 13 fatigue items. Higher scores indicate less fatigue. The minimally important differences (MIDs) are 7-8 points for the FACT-B [56] and 3-4 points for the FACIT-F subscales [57,58]. Both have shown a good test-retest reliability and internal consistency [52,53].

Stress
Perceived stress was assessed using the 10-item version of the Perceived Stress Scale (PSS) [59], rated for the past month on a 5-point rating scale (0 = "never", 1 = "almost never", 2 = "sometimes", 3 = "fairly often", 4 = "very often"). For the summed items (range from 0 to 40), a higher total score indicates greater stress, without a definite cutoff, because it is not a diagnostic instrument. The PSS has a good internal consistency and construct validity [60]. Perceived stress was assessed at weeks 0, 5, 17, and 29.
2.6.6. Treatment Expectation at Week 0 Treatment expectation was measured via a 100-mm visual analogue scale (VAS) for expectations concerning treatment as "expecting the treatment to be not successful at all" and "expecting the treatment to be extremely successful". All participants were asked before randomization to rate their expectations regarding both interventions separately. Only the expectation concerning the received treatment was used in the analysis.

Adverse Events
Adverse events were recorded by the acupuncturist or the psychologist, respectively, during each intervention using a standardized registration sheet. Furthermore, adverse events were assessed by enquiring self-reported adverse events of participants by week 5 and week 17 with open-ended questions. If the same adverse event was reported by the participant and the researcher, it was only accounted for once. Regardless of whether undesirable experiences during the course of the study were associated with the intervention or not, they were registered as adverse events. Serious adverse events were defined as: (1) death, (2) life-threatening situations, (3) hospitalization, (4) disability or permanent damage, (5) congenital anomaly/birth defect, or (6) the need for medical or surgical intervention [61,62]. All other adverse events were regarded as non-serious. If the licensed acupuncturist or psychologist suspected a causal relationship between the intervention and the adverse event, they were classified as intervention-related. Criteria for considering an adverse event to be intervention-related were: the specific adverse event had been previously reported in association with the intervention; it was clearly temporally related to the intervention; it occurred repeatedly in the intervention arm; and/or a causal relationship was physiologically plausible. The patient's assessment was also taken into account in the evaluation. If patients dropped out due to adverse events, this was noted.

Sample Size Calculation
An a priori power analysis with an estimated effect size f of 0.315 [63] (Cohen's d = 0.63), a 5% significance level, and 95% power for a mixed-model analysis of variance (ANOVA) with within-and between-subject interaction was performed. The analysis revealed that a total of 36 patients would be needed in order to detect an effect on the primary outcome. To compensate for a dropout rate of up to 30%, we planned to include a total of 52 patients in the study.

Statistical Analysis
All randomized patients were included in the analysis, regardless of whether they adhered to the allocated treatment and whether they had missing data. Missing data at all follow-up timepoints were multiply imputed. By using a multiple imputation technique, 20 additional datasets were generated using iterations based on multivariate regression models of sociodemographic and clinical parameters of baseline and outcome values. All analyses were based on the intention-to-treat sample and conducted with the Statistical Package for Social Sciences software (IBM SPSS Statistics for Windows, release 25.0; IBM Corporation, Armonk, NY, USA). A 2 (time: pre/post) × 2 (group: auricular acupuncture/psychoeducation) mixed-model ANOVA with treatment expectation as a covariate and sleep quality (total PSQI score) after treatment (week 5) as a dependent variable was conducted to test our primary hypothesis. Accordingly, secondary outcomes were evaluated in the same exploratory manner. To test moderation by the treatment expectation, the three-way interaction was also regarded. If a significant three-way interaction was found, a regression analysis with simple slope analyses was conducted. The non-parametric Fisher's exact test was used in an exploratory fashion to test whether there was a significant difference between the good sleeper/bad sleeper groups (PSQI ≥ 8) at week 5, and on the clinically significant improvement of 3 points by the intervention group. A p-value < 0.05 was considered significant, and partial eta squared (η 2 p ) was reported as an effect size estimator. According to Cohen, an η 2 p of 0.01 is considered a small, of 0.06 a medium, and of 0.14 a large effect [64].

Participant Characteristics
A total of 93 patients were assessed for eligibility ( Figure 2); 52 patients fulfilled the inclusion criteria, and were randomized to either auricular acupuncture (n = 26) or psychoeducation (n = 26). For baseline demographics, see Table 2. All patients in the auricular acupuncture group attended at least 8 of the 10 appointments. Seventy-seven percent attended the one-time psychoeducation group. At week 5 and week 17, 94.2% of participants provided data, and 86.5% of participants at week 29.

Primary Outcome
A significant interaction effect with a medium effect size was found for condition × time on sleep quality (p = 0.031, η 2 p = 0.094). Therefore, the auricular acupuncture in comparison with the psychoeducation had a significant effect on sleep quality over time (Figure 3). There was no statistically significant three-way interaction of expectation × time × group (p = 0.120, η 2 p = 0.050), indicating no significant effect of treatment expectation. Therefore, we did not conduct further simple slope analyses. At week 5, 19.2% of the control group and 61.5% of the intervention group had a PSQI score below 8 (p = 0.004). A clinically significant improvement of 3 points or more was observed in 38.5% of the control group and 65.4% of the intervention group (p = 0.095).

Adverse Events
A total of 55 adverse events occurred in 20 patients during the study period. All adverse events occurred in the auricular acupuncture group. All in all, 3 patients (5.8%) reported 1, 7 patients (13.5%) reported 2, 5 patients (9.6%) reported 3, 2 patients (3.8%) reported 4, and 3 patients (5.8%) reported 5 adverse events. All adverse events were nonserious, and 39 (75%) of the events in 16 patients were judged to be intervention-related by the researchers (Table 4).

Discussion
This is the first trial to examine the effect of auricular acupuncture on subjective sleep quality in breast cancer survivors with insomnia. We found that the subjective sleep quality in the auricular acupuncture group compared to the psychoeducation group improved significantly over the study period (5 weeks), controlled for treatment expectation. Furthermore, auricular acupuncture improved stress, anxiety symptoms, and fatigue at week 5 compared to psychoeducation. No effect was found for depressive symptoms and quality of life at 5 weeks, or for any outcome at week 17 or week 29. Additionally, no effect on the inflammation parameter (IL-6 level) could be detected due to floor effects, as there was no elevation of IL-6 at baseline. No serious adverse events occurred during the study period.
In comparison to prior acupuncture trials concerning cancer survivors with insomnia, the mean global PSQI score declined to a similar degree [65], and even stronger than in a trial on electroacupuncture [66]. Therefore, our auricular acupuncture regime seems to have comparable effects to body acupuncture and electroacupuncture. While in our trial the semi-standardized auricular acupuncture was used, some prior trials used the standardized NADA (National Acupuncture Detoxification Association) protocol. In a randomized controlled trial, 59 insomnia disorder patients without cancer diagnosis were treated with either CBT-I or auricular acupuncture [67]. Significant between-group improvements were seen in favor of CBT-I. Both groups showed significant within-group post improvements, which were maintained after six months. Nevertheless, the authors con-cluded that auricular acupuncture cannot be considered an effective stand-alone treatment for insomnia disorders.
In our trial, three months after the intervention, the global PSQI score in the psychoeducation group further improved, while effects slightly diminished in the auricular acupuncture group, though still showing an improvement compared to baseline, with no significant differences between the groups (Table 3, Figure 3). The psychoeducation as well as the advice booklet for dealing with sleep disorders-which both include elements of CBT-I-might explain the improvement [23], though we did not query whether the booklet had been read. The first-line therapy for insomnia is CBT-I; unfortunately, certain factors complicate the access to insomnia treatment, including system barriers, clinician barriers, and patient barriers [15]. The system barrier of limited access to CBT-I and the patient barriers of time constraints and reluctance to engage in CBT-I were found to restrict the use of CBT-I, with only 5% effectively seeking treatment [68]. A broader range of evidence-based therapies to treat insomnia is needed in order to offer patient-centered supportive cancer care.
Therefore, for patients who reject CBT-I, auricular acupuncture might be an option [18], especially since fewer side effects are expected in comparison to pharmacological treatment.
All adverse events recorded were non-serious, and 39 (75%) of the events in 16 patients were classified by the investigators as being intervention-related. Non-serious adverse events reported at least twice included bruising, hunger pangs, flushing, heavy eyelids, fatigue, tenderness, and pain. To date, safety data on auricular acupuncture in cancer patients are scarce. In a recent trial, ear pain was the most common adverse event, and was reported by 18.9% of patients [34].
Additionally, the adherence to auricular acupuncture in our trial was high, with all 26 patients attending at least 8 of the 10 acupuncture group therapies, indicating a broad acceptance. Part of the effect of auricular acupuncture is attributed to a stimulation of the auricular branch of the vagus nerve, increasing parasympathetic activity by simultaneously decreasing sympathetic nervous system activity [69,70]; this may lead to improvements of stress-related symptoms. Alleviating comorbid conditions seems to influence patients' perception of their ability to respond to acupuncture [24]. We took comorbid conditions such as hot flushes and pain into account using a semi-structured auricular acupuncture concept in order to choose additional individual acupuncture points. While many patients benefited from this regime, there were certain non-responders. The combination of acupuncture and sleep hygiene practice could enhance the treatment effect [24], and should be examined in future trials.
We were further interested in whether auricular acupuncture influences other health parameters, because breast cancer survivors who suffer from sleep disturbance often also experience fatigue and depression [71]. While we found a significant improvement in anxiety symptoms, fatigue, and stress at week 5, we did not find a significant improvement for depression or quality of life. Only 20% of the patients showed clinically relevant depressive symptoms in the HADS-D at baseline; therefore, it is difficult to detect improvements.
Despite several strengths, including sample size calculation and controlling for treatment expectation, there are a number of limitations to this study. First, the control group did not receive the recommended first-line therapy of CBT-I. Second, we did not use objective assessments for sleep quality, such as polysomnography. Third, we used a semistandardized protocol for auricular acupuncture; while being a limitation, this was also a strength, since we were able to take individual comorbid conditions into account [24].
In future trials, auricular acupuncture should be compared with CBT-I using a noninferiority design. To distinguish which patients benefit from which treatment, or to examine the effect of combining CBT-I with acupuncture, should be considered in future research.
Effective and easy-to-implement treatment options for cancer patients with insomnia need to be offered and integrated into the healthcare system.

Conclusions
Group auricular acupuncture might be an effective and safe intervention in treating sleep problems in breast cancer survivors in the short term, and may also reduce stress, anxiety, and fatigue. Long-term effects remain questionable.
Author Contributions: All authors contributed substantially to the conception and design, acquisition of data, or analysis and interpretation of data, as well as the drafting of the article or revising it critically for important intellectual content. Conceptualization and methodology, P. . The Karl and Veronica Carstens Foundation had no influence on the design and conduct of the study, the collection, management, analysis and interpretation of the data, the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.

Institutional Review Board Statement:
Prior to patient recruitment, approval was obtained from the ethics committee of the University of Duisburg-Essen (project number: 18-8214-BO, approval date 7 September 2018), and the study was registered at ClinicalTrials.gov (registration number: NCT03874598) prior to patient recruitment. The study was conducted in accordance with the tenets of the Declaration of Helsinki, and the report was written in conformance with the CONSORT 2010 Statement and the CONSORT harm extension.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: Data are available on request from the authors.