Trial design

This study was designed as a three-arm, parallel, RCT comparing ILM to sham ILM (SILM) or waitlist control (WLC) groups as a treatment for CRF among BCS. The trial design, recruitment/retention strategies, interventional approaches, and associated procedures have been published previously [14]. This trial received approval from the Institutional Review Board of Yueyang Hospital of Integrated Traditional Chinese Medicine and Western Medicine (IRB no. 2,018,029), and was performed in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines and the Standards for Reporting Interventions in Clinical Trials of Acupuncture (STRICTA).

Patients

This trial enrolled female breast cancer (stage I-III) survivors 18 years of age or older at a minimum of 12 weeks following primary treatment (including surgery, radiotherapy, and/or chemotherapy). Enrolled patients were those complaining of persistent fatigue that was moderate-to-severe (≥ 4 on a numerical rating scale of average fatigue) despite having the ability to rest. Patients with fatigue that was potentially attributable to a treatable condition such as hypothyroidism or anemia were excluded. Participants were excluded if they had received acupuncture for any indication in the previous four weeks, had received an acupuncture or drug test within the past six months, were pregnant or lactating, were diagnosed with severe mental illness, or had a systemic infection or infectious diseases. All patients provided written informed consent before participating in the study.

Randomization and masking

Potentially eligible patients were referred by oncology physicians who were contacted by research staff. In addition, trial-related information was posted on notice boards throughout Yueyang Hospital. Study clinicians met with interested patients to confirm that they met the criteria for eligibility. Those eligible patients that provided written informed consent were assigned to the ILM, SILM, or WLC groups at a 2:2:1 ratio, respectively. Randomization was achieved using a secure system with full allocation concealment, using permuted block randomization, with stratification by age and baseline level of fatigue. All statistical analyses were performed by statisticians blinded to the treatment assignments. The assessors and the statistician were blinded to treatment allocation throughout data collection and analysis. The study investigators, laser instrument operator, and participants were blinded to the treatment assignments between the real and sham laser moxibustion.

Procedures and interventions

Participants in ILM and SILM groups were treated in a private room at the appointed time and were instructed not to talk to each other. All patients in three groups maintained their usual treatment and self-care during the study.

Infrared laser moxibustion (ILM group)

Treatment of patients in the ILM and SILM groups was performed with SX10-C1 laser moxibustion devices (Shanghai Wonderful OptoElectrics Tech Co. Ltd., Shanghai, China). Four laser probes were simultaneously aligned with four selected acupoints (ST36 [bilateral], CV4, and CV6), with each of these points then being irradiated 2 cm from the surface of the skin for 20 min. For the ILM group, the output power level was 170 mW, the per-treatment energy density was 64.97 J/cm2, and the dose per treatment point was 203.91 J (eAppendix in Supplement Material). Treatment was performed twice per week for six weeks (12 total sessions) in all patients.

Sham infrared laser moxibustion (SILM group)

The treatment protocol for patients in the SILM group was identical to that for patients in the ILM group, except that there was no laser emission when the moxibustion instrument was turned on. As the infrared laser used in the ILM group was colorless, it was not visible to the operator or the patient such that the procedure was double-blind.

Waitlist control (WLC group)

No ILM or SILM treatment was provided to patients in the WLC group. These patients instead maintained standard treatment and self-care regimens, and did not initiate any additional treatments aimed at alleviating their CRF during the study period. After the completion of the study follow-up period, patients in the WLC group were provided with 10 real ILM treatments.

Outcomes

The primary outcome of the study was a change in the mean patient score on the Chinese version of the Brief Fatigue Inventory (BFI-C) [15] over time. Specifically, changes in the BFI-C scores were assessed at the end of treatment (week 6) and the end of follow-up (week 18) relative to baseline. The BFI includes three fatigue severity-related questions that are rated from 0 (no fatigue) to 10 (fatigue as bad as you can imagine), as well as 6 questions related to the degree to which fatigue interferes with particular functions, rated from 0 (does not interfere) to 10 (interferes completely) [16]. The mean score for these 9 items was the primary outcome variable for this study.

The secondary outcomes included patients’ scores on the Pittsburgh Sleep Quality Index (PSQI) [17, 18], the Hospital Anxiety and Depression Scale (HADS) [19], 10-item Perceived Stress Scale (PSS-10) [20], the Brief Pain Inventory (BPI) [21], and the Functional Assessment of Cancer Therapy-Breast (FACT-B) [22, 23]. These assessments were completed at baseline (week 0) and at weeks 3, 6, 12, and 18.

Moreover, the Acupuncture Expectancy Scale (AES) [24, 25] was also used to gauge participant response expectancy regarding laser moxibustion for patients in the ILM and SILM groups at baseline (week 0), week 3, and week 6 (end of treatment). A standardized adverse-event case report form was completed by research staff at each ILM/SILM treatment visit during the interventional period, and adverse events were also recorded during subsequent patient follow-up.

Statistical analysis

In our prior study, we found that ILM treatment yielded a significantly greater 0.60 standard deviation (SD) reduction in BFI-C scores at the end of a 4-week interventional period relative to baseline when compared to SILM treatment. With a two-sided alpha of 0.05, 45 subjects would thus be required in each of the ILM and SILM groups to detect a 0.60 SD reduction in BFI-C scores between these two groups at an 80% power level. By enrolling 45 subjects in each of the ILM and SILM groups and 22 subjects in the WLC group, it would be possible to detect a 0.71 SD change in BFI-C scores between the ILM and SILM or WLC groups at an 80% power level with a two-sided alpha of 0.05. Assuming that 20% of patients would be lost to follow-up, the study thus planned to enroll 140 patients that would be assigned to the ILM, SILM, and WLC groups at a 2:2:1 ratio. As the study used repeated measures and mixed-effects models to analyze the resultant data, providing higher statistical power than t-tests, this was a relatively conservative approach to study power calculation.

The study was conducted based on an intention-to-treat (ITT) principle. Baseline characteristics were compared among groups with one-way ANOVAs, Kruskal-Wallis and chi-square tests, or Fisher’s exact test. Changes in the BFI and their interactions with group assignment and time were evaluated using a linear mixed-effects model. Linear mixed-effects models were also used to assess the secondary PSQI, HADS, PSS-10, BPI, and FACT-B outcome measures. Linear mixed-effects models enable valid inferences based on the assumption that any missing follow-up data are missing at random. Missing values were thus not replaced for these analyses. Results were corrected for multiple comparisons via the Bonferroni correction, and a two-sided P < .05 was the threshold for significance in all analyses.