Qigong improves quality of life in women undergoing radiotherapy for breast cancer

Results of a randomized controlled trial




Radiotherapy may lead to side effects that undermine patients' quality of life (QOL). Although mind-body practices like qigong appear to improve QOL in cancer survivors, little is known about their benefits for patients who are receiving radiotherapy. Thus, in the current randomized controlled trial, the authors examined the efficacy of a qigong intervention on QOL in women with breast cancer during and after treatment.


Ninety-six women with breast cancer were recruited from a cancer center in Shanghai, China, and were randomized to a qigong group (N = 49) or a waitlist control group (N = 47). Women in the qigong group attended 5 weekly classes over 5 or 6 weeks of radiotherapy. QOL outcomes (ie, depressive symptoms, fatigue, sleep disturbance, and overall QOL) and cortisol slopes were assessed at baseline, during treatment, at the end of treatment, 1 month later, and 3 months later.


The mean age of the women was 46 years (range, 25-64 years). Seven percent of women had stage 0 disease, 25% had stage I disease, 40% had stage II disease, and 28% had stage III disease. Fifty-four percent of women underwent mastectomy. Multilevel analyses revealed that women in the qigong group reported less depressive symptoms over time than women in the control group (P = .05). Women who had elevated depressive symptoms at the start of radiotherapy reported less fatigue (P < .01) and better overall QOL (P < .05) in the qigong group compared with the control group, and these findings were clinically significant. No significant differences were observed for sleep disturbance or cortisol slopes.


The current results indicated that qigong may have therapeutic effects in the management of QOL among women who are receiving radiotherapy for breast cancer. Benefits were particularly evident for patients who had preintervention elevated levels of depressive symptoms. Cancer 2013. © 2013 American Cancer Society.


Despite incredible medical advances in cancer treatment, patients are faced with treatment side effects that undermine their quality of life (QOL). Patients with breast cancer are at particularly high risk, because they tend to receive multimodal treatment, including surgery, chemotherapy, and/or radiotherapy (RT), over an extended period,1, 2 leading to an accumulation of related morbidity. Side effects may include pain, fatigue, pulmonary symptoms, radiation burns, nausea, vomiting, insomnia, infertility, and sexual dysfunction among others.3 Not surprisingly, patients also are at risk of experiencing anxiety and depression.4-7 Psychological distress is problematic, because it contributes to hypothalamic-pituitary-adrenal (HPA) axis dysregulation,8, 9 which may have an impact on tumor progression, overall recovery, and possibly even survival.8, 10-12 A reliable measure of HPA axis dysregulation is to assess the diurnal cortisol pattern, a stress hormone and an end product of the HPA axis.13 Studies have revealed that both elevated levels of cortisol and flattened high cortisol levels or erratic diurnal cortisol fluctuations cause down-regulation of the immune response and are associated with worse survival in women with breast cancer.14, 15

Growing evidence suggests that stress-reduction techniques (eg, meditation, progressive relaxation, diaphragmatic breathing, guided imagery, and physical movements) improve QOL and reduce physiological stress (ie, HPA axis dysregulation).16-19 Although practicing these techniques individually appears to be successful in reducing psychological distress, a mind-body intervention, such as the ancient Chinese practice of qigong (“qi” energy flow, “gong” skill or achievement),20 that combines many of these techniques, may be particularly effective in addressing the multifaceted needs of patients with breast cancer who are undergoing active treatment. Through mind-body integration of specific movements, breathing techniques, meditation, and a mindful focus on the body, the primary objective of qigong is to manipulate the patient's energy (qi) for the benefit of their physical, psychological, and spiritual health. A recent review of clinical trials examining the effects of qigong in cancer patients who were receiving chemotherapy revealed generally positive findings.21, 22 For instance, in a sample of 81 patients with mixed cancer diagnoses, patients who participated in a 10-week qigong program experienced better overall QOL and less inflammation compared with those who received usual care.23 However, little is known about the benefits for patients who receive RT and the potential long-term effects, because none of the previous studies in cancer patients included follow-up assessments beyond the intervention period. Thus, we were interested in the effects of practicing qigong on QOL and biologic indices of stress (ie, cortisol function) during and after completing RT. Because QOL is a broad, multidimensional construct,20 we focused on the most frequently reported long-term QOL concerns in this population, including depressive symptoms, fatigue, and sleep disturbances, as well as overall QOL.24

We hypothesized that patients who practiced qigong would report better QOL (ie, fewer depressive symptoms, less fatigue and sleep disturbance, and improved overall QOL) compared with patients in a wait-list control (WLC) group during and after treatment. We also hypothesized that women in the qigong group would have less cortisol dysregulation, as reflected by a steeper diurnal cortisol slope and a smaller cortisol awakening response than the WLC group. Our secondary objective was to examine the effectiveness of the qigong program as a function of women's baseline depressive symptoms. A recent meta-analysis25 of 61 trials demonstrated that preintervention depressive symptoms/psychological distress moderated the efficacy of psychosocial/behavioral treatment for cancer patients, suggesting that psychosocial/behavioral interventions may be most beneficial for cancer patients with elevated depressive symptoms. Consequently, we hypothesized that patients who had high baseline depressive symptoms would benefit more from the qigong intervention that patients who had low baseline depressive symptoms.


Study Population

This trial was conducted at the Fudan University Shanghai Cancer Center (Shanghai, China) between 2005 and 2007 in collaboration with faculty from The University of Texas MD Anderson Cancer Center (Houston, Tex). Women with breast cancer (stages 0-III) who were aged ≥18 years; able to read, write, and speak Mandarin; had undergone breast surgery; and were scheduled to receive 5 to 6 weeks of RT were eligible to participate in the study. Patients who were unable to legally provide consent and those who reported prior regular qigong or tai chi practice (self-defined) within the past year were excluded from the study.


During pretreatment clinic visits, immediately after the treating radiation oncologist had briefly introduced the study to potentially eligible patients, research nurses approached patients, confirmed study eligibility, and obtained written informed consent before data collection. Before the start of RT and randomization, patients underwent a baseline (T1) assessment of questionnaires and provided saliva samples at 4 time points daily for 2 consecutive days to assess their cortisol rhythm according to procedures described elsewhere.13, 26 Participants who were receiving RT as outpatients were then assigned randomly to either the qigong group or the WLC group by a form of adaptive randomization called minimization,27 so that the groups were balanced evenly according to disease stage, age, time since diagnosis, type of surgery (mastectomy vs breast-conserving surgery), and prior chemotherapy. Participants who were receiving RT as inpatients were assigned to cohorts to minimize bias and contamination of the control group, because all inpatient participants resided on the same floor during treatment. Every inpatient participant in a given cohort was assigned to the same group as determined by a coin flip. Once everyone in the cohort had finished her treatment, the next cohort received the alternate condition. Participants in the qigong group followed the qigong program as described below. Women in the WLC group received the standard of care.

Participants in both conditions completed 4 additional assessments, including questionnaires and saliva sample collection in the middle of RT (T2), during the last week of RT (T3), and 1 month (T4) and 3 months (T5) after the end of RT. Patients were given a small gift (a $20 value) for completing each assessment. If the patients were not coming for a follow-up visit or were unable to travel, a questionnaire and saliva kit were mailed to them. Women who were assigned to the WLC group were given the option of attending the qigong program after they completed the last assessment, but no additional data were collected. The institutional review boards of both The University of Texas MD Anderson Cancer Center and Fudan University Shanghai Cancer Center approved the protocol.

Qigong Program

Women in the qigong group attended five 40-minute qigong classes each week during their 5 or 6 weeks of RT. All qigong sessions were taught by the same instructor, a government-licensed traditional Chinese medicine physician and qigong master who was trained at the Shanghai University of Traditional Chinese Medicine with 30 years of clinical experience. Session attendance ranged from 1 to 10 participants. Participants were given a DVD with a recording of the qigong program and some printed materials with similar instructions and were encouraged to practice qigong on their own on the days when they did not meet with the qigong master and after they had completed RT.

The program involved a modified version of Chinese medical qigong therapy developed by Guo Lin (Goulin New Qigong), often referred to as “walking qigong,”21 that is prescribed frequently in China for medical conditions, including cancer, and consisted of several breathing and moving exercises. After a brief course introduction, each session was separated into the following 3 parts: 1) The preparation exercise consisted of relaxation (gentle breathing and meditation; 4 minutes); synchronizing the breath with slow, shallow squatting movements (3 minutes); and synchronizing the breath with gentle arm movements in front of the abdomen (opening and closing of the dantian; 3-6 minutes). 2) For the main exercise, participants walked in a circle, synchronizing their breathing, arm movements, and steps and focusing on the movement of their body with the goal of calming their mind, relaxing various parts of the body and the mind, and revitalizing the “life-force” (ie, qi). This was done first slowly and then quickly using wind-like breathing (18 minutes). 3) Finally, the ending exercise consisted of breathing exercises, opening and closing of the dantian, and self-massage (3-6 minutes) practiced in a standing position.


All questionnaires were administered in Chinese using validated, translated versions of the original instruments.28-31 Symptoms of depression were assessed using the Center for Epidemiologic Studies Depression Scale (CES-D),28 a 20-item, self-report measure of depression that focuses on depressive feelings and behaviors over the past week. Higher scores on the CES-D represent worse symptoms, and a cutoff score ≥16 indicates “caseness,” warranting further psychological evaluation for clinical depression.

Fatigue was assessed using the Brief Fatigue Inventory (BFI),31 a 9-item questionnaire that asks participants to rate the severity of their fatigue at that moment and how much it interfered with their lives during the previous 24 hours. Higher scores on the BFI represent worse fatigue.

Sleep disturbances were assessed using the Pittsburgh Sleep Quality Index (PSQI),29 an 18-item, self-rated questionnaire that assesses sleep disturbances over the past month. The instrument includes 7 subscales: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medications, and daytime dysfunction. The current study reports on the total PSQI score. Higher scores represent greater sleep disturbances.

Overall QOL was assessed with the 28-item Functional Assessment of Cancer Therapy-General (FACT-G)30 questionnaire to assess health-related QOL, including physical, functional, social, and emotional well being dimensions. Research indicates that a total difference ≥7 in the FACT-G score is clinically significant.32

The cortisol rhythm also was assessed. To determine the circadian slope of cortisol, 4 saliva samples per day (waking, 45 minutes later, 8 hours later, and at bedtime) were obtained for 2 consecutive days.33 Levels of cortisol were determined using a time-resolved immunoassay with fluorescence detection.

Demographic and Medical Factors and Tracking Data

Some demographic items (eg, age, marital status) were included in the baseline questionnaires. Medical data were extracted from patients' charts. Tracking data were kept regarding study attrition, qigong class attendance, and frequency of home practice. Participants in the qigong group completed a brief evaluation of the class and their home practice each week when they attended the qigong classes and with their questionnaires at later assessments (T4 and T5).

Data Analysis

The primary outcome was the 1-month QOL indices (ie, fewer depressive symptoms, less fatigue and sleep disturbance, and improved overall QOL). A sample size of approximately 50 patients per group provided 80% power (2-sided alpha; P = .05) to detect effect sizes between .56 and .63 assuming 10% attrition. We calculated descriptive statistics and examined whether demographic and medical factors (eg, age, income, disease stage, time since surgery, and type of surgery) were related significantly to outcomes at P < .05; if so, then these factors were included as control variables in the main analyses along with inpatient/outpatient status. Main analyses were performed with multilevel modeling analyses using Proc MIXED in the SAS statistical software package (version 9.2.2; SAS Institute, Inc., Cary, NC) to account for the nested data structure of our repeated-measure design for each outcome variable. To evaluate the general effectiveness of the qigong intervention, we examined the group × time interaction effect while controlling for main effects and baseline scores of the outcome variable. For significant group × time interactions, we compared the least squared means (LSM) for each group at each time point using a general linear model analysis, controlling for baseline levels of the outcome variable and illustrating the interaction by plotting the LSM. Because of non-normal distribution of cortisol data, the raw levels were log-transformed. Cortisol circadian slopes were calculated by regressing log-transformed cortisol levels on the saliva collecting time (hours after waking up in the morning) for each participant at each assessment point. The cortisol awakening response was calculated by examining the difference between cortisol levels at wakening and 45 minutes later. To evaluate whether the intervention was more effective for participants with high scores (ie, scores 1 standard deviation above the sample mean) for baseline depressive symptoms, we examined the group × time × baseline depression interaction effect for QOL outcomes. Because <5 comparisons were performed, and comparisons involved complementary constructs, we did not use a multiple comparison adjustment.34, 35 Furthermore, because our analyses involved the testing of interaction terms, all significance tests were highly conservative.36


Baseline Characteristics of the Sample

We approached 123 eligible women, and 100 consented to participate in the study. Fourteen 14 patients failed to provide a reason for study participation, others indicated that a lack of interest (n = 3), physical symptoms (n = 3), spousal concerns (n = 2), or time constraints (n = 1) prevented their participation. After randomization and before the start of RT, 1 patient became ineligible (discontinued RT), and 3 patients withdrew from the study (lack of interest, spouse's wishes, and medical complications), leaving 49 patients in the qigong group and 47 patients in the WLC group. All participants completed all assessments except 1 patient in the WLC group, who died after the T4 assessment (Fig. 1).

Figure 1.

This is a Consolidated Standards of Reporting Trials (CONSORT) flow diagram. T1 was the baseline assessment of questionnaires and cortisol saliva levels, and cortisol levels also were measured in saliva samples that were obtained at 4 time points per day for 2 consecutive days as follows: T2, during the middle of radiotherapy (RT); T3, during the last week of RT; and (T4) 1 month and (T5) and 3 months after the end of RT.

Baseline characteristics of the participants assigned to each condition were analyzed using chi-square tests for categorical variables (eg, inpatient vs outpatient, income, type of surgery, prior chemotherapy) and t tests for continuous variables (eg, age, time since surgery). There were no significant differences in patient characteristics between groups (see Table 1). The sample mean ± standard deviation of CES-D score was 12.64 ± 9.01, and 39.2% of the sample met the CES-D criterion for caseness. It is noteworthy that there were no differences between groups at baseline in any of the outcome variables (depressive mood, fatigue, sleep disturbance, overall QOL, and cortisol slope).

Table 1. Participant Demographics and Clinical Characteristics by Group
 No. of Patients (%) 
VariableQigong, n = 49WLC, n = 47P
  • Abbreviations: RT, radiotherapy; SD, standard deviation; WLC, wait-list control receiving usual care.

  • a

    Income was converted to US dollars from Chinese Yuan.

Age: Mean ±SD [range], y45.3±6.3 [29-58]44.7±9.7 [25-62].73
Marital status: Married46 (93.9)40 (87).25
Highest level of education: Some college or higher26 (55)23 (50).17
Income, $USa   
 <15002 (4.1)4 (8.7).12
 1500-790020 (40.8)21 (45.7) 
 8000-15,80010 (20.4)12 (26.1) 
 >15,8007 (14.3)4 (8.7) 
 Refused to answer10 (20.4)5 (10.9) 
 02 (4.3)3 (6.7).20
 I8 (17.4)14 (31.1) 
 II17 (34.7)17 (36.2) 
 III17 (34.7)7 (14.9) 
Time since surgery: Mean±SD [range], wk16.5±5.5 [2.14-23.86]14.7±5.7 [0-25.1].24
Surgery type   
 Mastectomy18 (36.7)26 (56.5).07
 Lumpectomy31 (63.3)20 (43.5 
No. of RT fractions   
 2539 (78.3)35 (77.8).90
 309 (19.6)10 (22.2) 
 351 (2.2)0 (0) 

Adherence to Qigong Program

Adherence to the qigong program was high, and 30.4% of women attended 100% of sessions, 65.2% attended ≥80% of sessions, 78.3% attended ≥50% of sessions, and only 13% attended <20% of sessions (mean ± standard deviation, 22.0 ± 6.1 sessions; range, 5-30 sessions).

Main Effect of Intervention: Depressive symptoms

Multilevel modeling analyses revealed a significant main effect of time (F[4,367] = 10.91; P = .001), indicating that depression scores decreased over time. However, this effect was moderated by group (F[3,281] = 2.62; P = .05), indicating significant group differences in changes in depression over time. Figure 2 illustrates the reduction in depressive symptoms over time for the qigong group but no changes in the WLC group. Follow-up analyses at specific time points revealed marginally significant group difference at the T4 assessment (F[1,92] = 2.97; P = .09; adjusted means: qigong vs WLC, 10.01 vs 12.14). No significant differences between groups were observed for fatigue, overall QOL, sleep disturbance, cortisol circadian slope, or cortisol awakening response (for means by group over time for each variable, see Table 2).

Figure 2.

This chart illustrates the least squared means (adjusted for baseline level) of a multilevel modeling analysis for depressive symptoms over time. Higher scores represent greater symptoms. CES-D indicates the Center for Epidemiologic Studies Depression Scale; RT, radiotherapy.

Table 2. Raw Means of Outcome Measures by Group at Each Assessment Point
 Assessment Point
 BaselineMiddle of RTLast Week of RT1 Month Post-RT3 Months Post-RT
Outcome MeasureMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSD
  1. Abbreviations: BFI, Brief Fatigue Inventory; CAR, cortisol awakening response; CES-D, Center for Epidemiologic Studies Depression Scale; FACT-G, Functional Assessment of Cancer Therapy-General; PSQI, Pittsburgh Sleep Quality Index; RT, radiotherapy; SD, standard deviation; Slope, cortisol circadian slope; WLC, wait-list control.


Depression As a Moderator of Intervention Effectiveness

Depressive symptoms

There was a significant group × time × baseline depression interaction effect (F[4,368] = 3.98; P = .004) for depressive symptoms. We decomposed the interaction according to high and low (mean ± 1 standard deviation) baseline depressive symptoms and plotted the LSM for each group across time (Fig. 3a). There were no significant differences between groups for those with low baseline depressive symptoms at any assessment point. Among the women with high baseline depressive symptoms, there were marginally significant difference between groups at T3 (F[1,92] = 3.78; P = .06; qigong vs WLC, 16.36 vs 23.6), and significant there were differences between groups at T4 (F[1,92] = 13.32; P < .001; qigong vs WLC, 15.5 vs 29.1) and at T5 (F[1,90] = 6.48; P < .05; qigong vs WLC, 12.74 vs 26.51).

Figure 3.

The least squared means (adjusted for baseline level) of a multilevel modeling analysis are illustrated by group for those with high and low baseline depressive symptoms (mean ± 1 standard deviation) over time for (a) depressive symptoms on the Center for Epidemiologic Studies Depression Scale (CES-D), (b) fatigue measured with the Brief Fatigue Inventory (BFI), and (c) overall quality of life measured on the Functional Assessment of Cancer Therapy-General (FACT-G) quality-of-life questionnaire. Higher scores represent greater symptoms and fatigue and better overall quality of life. RT indicates radiotherapy.


Controlling for stage at diagnosis (which was associated with fatigue levels), there was a significant group × time × baseline depression interaction effect (F[3,275] = 4.25; P = .01) for fatigue. To illustrate the overall effect of the significant interaction, we decomposed the interaction according to high and low baseline depressive symptoms and plotted the LSM for each group across time (Fig. 3b). There were no significant differences between groups for those with low baseline depressive symptoms at any assessment point. For women with high baseline depressive symptoms, there were significant differences between groups at T4 (F[1,90] = 5.36; P < .05; qigong vs WLC, 2.93 vs 4.19).

Overall quality of life

Controlling for disease stage at diagnosis and type of resection (which were associated with overall QOL), there was a significant group × time × baseline depression interaction effect (F[3,275] = 3.03; P = .02) on overall QOL. To illustrate the overall effect of the significant interaction, we decomposed the interaction according to high and low baseline depressive symptoms and plotted the LSM for each group across time (Fig. 3c). There were no significant differences between groups for those with low baseline depressive symptoms at any assessment point. For women with high baseline depressive symptoms, there were significant differences between groups at T4 (F[1,90] = 7.71; P < .01; qigong vs WLC, 82.44 vs 66.31). It is noteworthy that the differences between groups also were clinically significant (between-group difference, 7 points).32 At T5, there was a clinically significant (but statistically nonsignificant) difference (qigong vs WLC, 85.55 vs 78.11) for women who had high baseline depressive symptoms.

Sleep disturbances and cortisol outcomes

Baseline depressive symptoms did not significantly moderate the group × time interaction effect on sleep disturbances, cortisol circadian slope, or cortisol awakening response.

Treatment adherence

Class attendance and home practice were not associated significantly with any of the outcomes.


The current randomized controlled trial indicated that qigong may reduce depressive symptoms in women with breast cancer who are receiving RT. Qigong may be particularly helpful for women who report elevated levels of depressive symptoms at RT initiation. When baseline symptoms of depression were taken into account, group differences revealed lower depressive symptoms and fatigue and better overall QOL for the qigong group versus the WLC group. This finding may have important implications given the association between depression and survival in prospective studies.10 The observed benefits exceeded the threshold for clinically significant differences in overall QOL32 at 1 month and 3 months after RT. Generally, observed significant differences in outcomes (depressive symptoms, fatigue, and overall QOL) emerged after treatment but not during treatment, suggesting that qigong may prevent delayed symptom burden or expedite the RT recovery process for women who have elevated levels of depressive symptoms at the start of RT.

This research supports the promising findings of previously reported qigong trials22, 23; however, to our knowledge, this is the first trial to enroll women who were actively receiving RT and to include a follow-up period. Because no adverse events were reported in our study participants, qigong appeared to be a safe therapy for patients who were actively receiving treatment, and it was well accepted: 65% of women attended ≥80% of the qigong sessions. Our findings also support the results from a recent, large meta-analysis25 suggesting that behavioral interventions are most efficacious for participants who have increased baseline levels of depressive symptoms or distress. Consequently, although our trial demonstrated reduced depressive symptoms over time for the qigong group with no changes in the WLC group, the findings suggest that patients with depressive symptoms may benefit the most from a qigong intervention. Nevertheless, these findings are too preliminary to offer clinical recommendations, and further research is needed to examine the differential effects for all comers versus at-risk patients and to compare the benefits of qigong with other mind-body interventions (eg, yoga37). Future research also is needed to uncover the psychosocial and biologic mechanisms by which qigong affects symptom relief. We examined 1 mechanism, cortisol rhythm, which yielded no significant differences between groups.

Our study had some limitations. We did not include an active control intervention. Thus, we cannot rule out the possibility that the effects of the intervention were influenced by patients' expectations of the qigong practice or the attention and social support participants in the qigong group received over the 5-week or 6-week intervention period. We are currently addressing these concerns in a 3-arm, single-blinded qigong trial. Also, the homogenous nature of our sample, which was specifically Chinese women at a single site in Shanghai, China, limits the generalizability of our results. Therefore, it would be prudent to further explore the benefits of qigong in culturally and ethnically diverse samples of patients with breast cancer and other cancer types, including nonsex-specific disease sites. The finding of the benefits of qigong based on pre-RT depressive symptoms needs to be interpreted with caution, because our patients were not recruited based on depressive symptoms. Finally, we followed participants for only 12 weeks after RT; thus, the long-term effectiveness of qigong among patients with breast cancer remains to be determined.

In conclusion, the results from this randomized controlled trial indicate that qigong may lead to reductions in depressive symptoms over time and may be especially useful for women who report high symptoms of depression at baseline. Studies involving a blinded design with an active control group, selecting for high levels of baseline depressive symptoms, and longer term follow-up will be necessary to examine the generalizability of our findings and to deepen our understanding of this promising therapeutic approach.


We thank Drs. Peiying Yang, Joseph Chiang, and Qingyi Wei for all of their support with language, culture, and politics.


Support for this study was provided in part by the United States National Cancer Institute (NCI) (grants R21CA108084 and U19CA121503; principal investigator, Lorenzo Cohen) and by NCI Cancer Center Support grant CA016672.


The authors made no disclosures.