Breast Cancer Research and Treatment

https://doi.org/10.1007/s10549-019-05278-w

REVIEW

Yoga has a solid effect on cancer‑related fatigue in patients

with breast cancer: a meta‑analysis
Bei Dong1,2 · Congyan Xie1,2 · Xiuchen Jing1 · Lu Lin1,2 · Li Tian1,2

Received: 30 April 2019 / Accepted: 10 May 2019

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
Purpose This study was designed to critically evaluate the effect of yoga on cancer-related fatigue in patients with breast
cancer.
Methods Eight databases (Cochrane Library, PubMed, Ovid-Medline, Web of Science, CBM, Wanfang, VIP, and CNKI)
were systematically reviewed from inception to January 2019 for randomized controlled trials (RCTs). Two reviewers criti-
cally and independently assessed the risk of bias using Cochrane Collaboration criteria and extracted correlated data using
the designed form. All analyses were performed with Review Manager 5.3.
Results A total of 17 qualified studies that included 2183 patients (yoga: 1112, control: 1071) were included in the meta-
analysis. Yoga had a large effect on fatigue in post-treatment breast cancer patients and had a small effect on intra-treatment
patients. The meta-analysis also indicated that supervised yoga class had a significant effect on CRF; the six-week program
had a moderate beneficial effect while the 60/90 min/session supervised yoga class and the eight-week program demonstrated
a large effect on fatigue in patients with breast cancer. Yoga could markedly mitigate the physical fatigue in breast cancer
patients, had a medium impact on cognitive fatigue, and manifested a small effect on mental fatigue. Eight studies reported
the adverse events, whereas ten studies did not.
Conclusions Yoga can be considered as an alternative therapy for relieving fatigue in breast cancer patients who have com-
pleted treatment or are undergoing anti-cancer treatment.

Keywords Yoga · Fatigue · Breast cancer · Meta-analysis

Introduction identified and 0.52 million cases of deaths caused by breast

cancer occurred in 2012 globally [1]. The incidence of breast
Breast cancer is the most common cancer in women around cancer is predicted to increase to 85 per 100,000 women by
the world, accounting for 25.1% of all cancers. It was esti- 2021 [2]. Increased incidence and improved treatments have
mated that 1.67 million new cases of breast cancer were made breast cancer survivors more prevalent in the world
[1]. Cancer-related fatigue (CRF) is defined as a distressing,
persistent, subjective sense of physical, emotional, and/or
Bei Dong and Congyan Xie have contributed equally to this work.
cognitive tiredness or exhaustion related to cancer or can-
Li Tian is the first corresponding author and Lu Lin is the co- cer treatment that is not proportional to recent activity and
corresponding author interferes with usual functioning. It is one of the most com-
mon side effects in patients with cancer [3]. Women with
* Lu Lin
linlu@suda.edu.cn breast cancer may be already experiencing fatigue before
anti-cancer treatment. 60 to 90% of women with breast can-
* Li Tian
tianlisz@suda.edu.cn cer may experience fatigue at the beginning of treatment [4].
Approximately one in four breast cancer survivors (BCS)
1
The First Affiliated Hospital of Soochow University/School suffer from severe fatigue [5]. It is obvious that the number
of Nursing, Soochow University, No. 188 Shizi Road, of breast cancer patients with fatigue is enormous and the
Suzhou 215006, People’s Republic of China
management of CRF is urgent.
2
School of Nursing, Soochow University, Suzhou 215006,
People’s Republic of China

13
Vol.:(0123456789)
 Breast Cancer Research and Treatment

Yoga is an ancient oriental fitness method, which origi- Outcomes

nated in India. Traditional Indian yoga involves several
fields, including ethical thought, body posture, and spiritual Trials regarding fatigue as a primary or secondary outcome
practice, aiming at achieving the unity of thought, body, and containing fatigue scores that could be extracted were
mental health, and self-consciousness [6]. Many studies included.
have reported the effect of yoga on CRF in breast cancer
patients, some of which had positive outcomes, while others Studies
did not. In addition, the relationship between the benefits of
yoga and the type, duration, length, and frequency of yoga Only RCTs were eligible.
is still unspecific and ambiguous. Therefore, the purpose of
this meta-analysis was to critically assess the effect of yoga Data extraction
interventions on CRF in patients with breast cancer.
The general information of the studies, the characteristics of
the patients, the characteristics of the intervention and the
Methods control groups (form of intervention, type of yoga, dura-
tion and frequency of intervention, length of program, and
The meta-analysis was performed following the PRISMA adherence or contamination), the outcome measures, and the
guidelines for systematic reviews and meta-analyses [7]. adverse events were independently extracted by two review-
ers using a pre-designed information sheet.
Searching strategies
Risk of bias assessment
Eight databases, namely the Cochrane Central Register of
Controlled Trials, PubMed, Ovid-Medline, Web of Science, Two reviewers independently evaluated the risk of bias in the
China Biology Medicine (CBM), Wanfang, VIP, and China included RCTs using the Cochrane assessment tool, which
National Knowledge Infrastructure (CNKI) databases, were consists of the following seven domains: “adequate sequence
systematically searched from inception through January generation, allocation concealment, blinding of participants
2019 for relevant randomized controlled trials (RCTs) with- and personnel, blinding of outcome assessment, incomplete
out language restrictions. These searches were performed outcome data, selective reporting, and other bias” [7]. Each
using the following keywords: “Breast Neoplasm,” “Breast question can be rated as follows: yes (+), low risk of bias;
Tumor,” “Breast Cancer,” “Breast Carcinoma,” “Mammary unclear (?), unclear risk of bias; no (−), high risk of bias.
Cancer,” “Mammary Carcinoma,” “Mammary Neoplasm,”
“Mammary Tumor,” “Fatigue,” “Asthenia,” “Lassitude,” Data analysis
“Yoga,” “Yogic,” “Asana,”. Two reviewers first screened the
literature by scanning the titles and abstracts and then read The meta-analysis was performed using Review Manager
the full text of potentially eligible trials to decide whether Software (version 5.3). Clinical heterogeneity (variability
they should be included in the meta-analysis. The search in the participants, interventions, and outcomes) and meth-
strategy for the database PubMed is shown in the Appen- odological heterogeneity (variability in the study design and
dix as an example. Additionally, further potentially relevant risk of bias) were assessed first. Statistical heterogeneity is a
papers were searched using the reference lists of the identi- consequence of clinical or methodological diversity or both
fied articles. among the studies [7]. If moderate clinical heterogeneity
was identified, subgroup analyses were conducted on sta-
Inclusion criteria tus of treatment, type of yoga, form of intervention, and
dimension of fatigue if there were at least two studies on a
Participants stratum, considering that these variables might have influ-
ences on the fatigue outcomes. I2 was used to measure the
Studies including adult patients (≥ 18 years) who were diag- statistical heterogeneity among the trials in each analysis. If
nosed with breast cancer, regardless of cancer stage and cur- P > 0.1 and I2 < 50%, a fixed-effects model was adopted for
rent treatment, were eligible. the analysis due to the homogeneity of the studies; if P < 0.1
and I2 ≥ 50%, then a random-effects model was adopted. If
Interventions and controls P < 0.1 and the sources of heterogeneity were unknown, a
descriptive analysis was conducted without a meta-analysis.
Trials that compared any type of yoga with any type of con- For continuous data, the weighted mean difference (WMD)
trol group were included. and 95% confidence interval (CI) were determined for the

13
Breast Cancer Research and Treatment

individual trials. The standardized mean difference (SMD) duplicated (Fig. 1). After an initial review of the title and
was used if the outcome assessment tools were different. abstract, 583 records were excluded due to obvious non-
Reporting and publication bias were investigated by visu- conformity to the inclusion criteria. Of the remaining 25
ally examining the degree of asymmetry of a funnel plot. A articles, seven articles were excluded after reading the litera-
sensitivity analysis was performed in light of the fact that ture and critical appraisal, and thus 18 articles were included
some of the trials (e.g., one with a larger sample) might in systematic review. One article [22] was excluded from
impact the study results. Sensitivity analysis was used to further analysis because it did not provide the total score of
explore the effects of the fixed-effects or random-effects fatigue. Therefore, a total of 17 articles were included in the
model analyses for outcomes with heterogeneity and the final meta-analysis [8–21, 23–25].
effects of any assumptions.
Characteristics of the included trials

Results The characteristics of the patients, treatments, controls, out-

come measures, and adverse events are shown in Table 1.
Literature search
Participants
A total of 690 records identified through database searching
and three additional records identified through other sources Among the 18 studies included in systematic review,
were retrieved in the literature search, and 85 of them were the mean age of the participants ranged from 47.38 to

Fig. 1  Flow chart diagram of

trial identification and selection

13
 Table 1  Studies included in the systematic review
Author, year Sample Mean age Ethnicity Cancer stage Current treat- Treatment group: form of Control group: Fatigue scale Adverse events
(± SD) ment intervention, duration, frequency, intervention,

13
Gender length of program, program length,
contamination

Moadel, 2007 YG 84 YG %African I–IV Surgery/chemo- Yoga class(Hatha yoga) + home- Standard care, FACIT-fatigue Not reported
[8] CG 44 55.11 ± 10.07 American therapy/anties- based yoga, 90 min + unclear, 1 12 weeks, not
CG 54.23 ± 9.81 YG 42% trogen therapy/ time/week +daily, 12 weeks reported
Female CG 43% radiation
%Hispanic treatment
YG 30%
CG 34%
%Non-Hispanic
white
YG 22%
CG 23%
Others
Vadiraja, 2009 YG 44 Not reported Not reported I–III Radiotherapy Integrated yoga class + self- Supportive EORTC-QOL- Not reported
[9] CG 44 Female practice in the hospi- counseling, C30
tal,60 min + unclear, 3 times/ 6 weeks, not
week + unclear, 6 weeks reported
Danhauer, 2009 YG 22 YG 54.3 ± 9.6 %Non-Hispanic DCIS Chemotherapy/ Restorative yoga class, 75 min, 1 Wait-list, FACT-fatigue None
[10] CG 22 CD 57.2 ± 10.2 White I–IV radiation time/week, 10 weeks 10 weeks, not
Female YG 86.4% therapy reported
CG 90.9%
Banasik, 2011 YG 9 YG 63.33 ± 6.9 Caucasian II–IV Post-treatment Iyengar yoga class, 90 min, 2 Non-interven- Fatigue Likert not reported
[11] CG 9 CG 62.4 ± 7.3 times/week, 8 weeks tion, 8 weeks, Scale
Female not reported
Littman, 2012 YG 32 YG 60.6 ± 7.1 %Non-Hispanic In situ I–III Post-treatment Yoga class (viniyoga) Wait-list, FACIT-fatigue None
[12] CG 31 CG 58.2 ± 8.8 White +Home-based yoga, 24 weeks, not
Female YG 93.7% 65–85 min + 20–30 min, 5 reported
CG 93.5% times/week(yoga class + home-
based yoga), 24 weeks
Bower, 2012 YG 16 YG 54.4 ± 5.7 %White 0–II Post-treatment Iyengar yoga class, 90 min, 2 Health educa- FSI YG: back spasm1
[13] CG 15 CG 53.3 ± 4.9 YG 94% times/week, 12 weeks tion classes,
Female CG 80% 12 weeks, not
reported
Loudon, 2014 YG 12 YG 55.1 ± 2.5 Not reported DCIS Chemotherapy/ Yoga class + home-based yoga Usual self-care, VAS None
[14] CG 11 CG 60.5 ± 3.6 I–III radiotherapy sessions (Satyananda yoga), 8 weeks, not
Female 90 min + 45 min, 1 time/ reported
week + daily, 8 weeks
Breast Cancer Research and Treatment
 Table 1  (continued)
Author, year Sample Mean age Ethnicity Cancer stage Current treat- Treatment group: form of Control group: Fatigue scale Adverse events
(± SD) ment intervention, duration, frequency, intervention,
Gender length of program, program length,
contamination

Wang, 2014 [15] YG 40 Not reported Not reported Not reported Chemotherapy Yoga class (centralized interven- Usual care, CFS Not reported
CG 42 Female tion for inpatients/decentral- 16 weeks, not
ized intervention for outpa- reported
tients during the intermission
of chemotherapy), 50 min, 4
Breast Cancer Research and Treatment

times/week, 16 weeks
Taso, 2014 [16] YG 30 49.27 ± 10.23 Not reported I–III Chemotherapy/ Anusara yoga class, 60 min, Standard care, BFI Not reported
CG 30 Female antihormone 2 times/week, 8 weeks 8 weeks, not
therapy reported
Cramer, 2015 YG 19 YG 48.3 ± 4.8 Not reported I–III Anti-hormonal Yoga class(Hatha yoga) + home- Usual care, FACIT-fatigue YG: panic attack
[17] CG 21 CG 50.0 ± 6.7 therapy based yoga, 90 min + unclear 1 12 weeks, not and foot pain2,
Female time/week + daily, 12 weeks reported transient muscle
soreness3, uni-
lateral hip pain1
CG: sciatica1,
port pain1,
elbow pain1,
knee pain2,
panic attacks1
Vardar Yağlı, YG 19 YG 49.89 ± 4.65 Not reported I–II Chemotherapy/ Aerobic exercise added to yoga,Aerobic exercise, FSS/EORTC- Not reported
2015 [18] CG 21 CG 47.38 ± 7.57 radiotherapy 60 min, 3 times/week, 6 weeks 6 weeks, not QOL-C30
Not reported reported
Vardar Yagli, YG 10 YG 68.58 ± 6.17 Not reported I–II Post-treatment Classical yoga class, not Exercise pro- VAS Not reported
2015 [19] CG 10 CG 68.88 ± 2.93 reported, 2 times/week, gram, 4 weeks,
Not reported 4 weeks not reported
Lötzke, 2016 YG 45 YG 51.0 ± 11.0 Not reported I–III Chemotherapy/ Yoga class (Iyengar Conventional CFS-D/EORTC- Not reported
[20] CG 47 CG 51.4 ± 11.1 endocrine Yoga) + home-based yoga, physical exer- QLQ-C30
Female therapy/radia- 60 min + 20 min, 1 time/ cise, 12 weeks,
tion week + 2 times/week, 12 weeks not reported
Stan, 2016 [21] YG 18 YG 61.4 ± 7.0 Not reported 0–II Chemotherapy/ Home-based yoga, 90 min, 3–5 Strengthening MFSI-SF YG: Flu-like
CG 16 CG 63.0 ± 9.3 radiation/endo- times/week, 12 weeks exercises, symptom 3 Leg
Female crine therapy/ 12 weeks, not cramps 1 Swell-
mastectomy reported ing of the hand 1
Stomach pain 1
Shoulder pain 1
Side pain 2 CG:
Flu-like symp-
tom 2 Arm pain
1 De Quervain’s
tenosynovitis 1

13
 Table 1  (continued)
Author, year Sample Mean age Ethnicity Cancer stage Current treat- Treatment group: form of Control group: Fatigue scale Adverse events
(± SD) ment intervention, duration, frequency, intervention,

13
Gender length of program, program length,
contamination

Jin, 2017 [22] YG 50 Unclear Not reported Not reported Chemotherapy Yoga added to usual Usual care, CFS Not reported
CG 50 care(centralized intervention 16 weeks, not
for inpatients/decentralized reported
intervention for outpatients
during the intermission of
chemotherapy), 60 min (cen-
tralized intervention), 3 times/
week (centralized interven-
tion), 16 weeks
Zeng, 2017 [23] YG 24 Not reported Not reported Not reported Chemotherapy Yoga added to usual care, CG1: usual care, CFS Not reported
CG1 23 40 min, once every two days, 4 weeks, not
CG2 20 4 weeks reported CG2:
CG3 22 music relaxa-
tion training
add to usual
care, 4 weeks,
not reported
CG3: yoga
combined with
music relaxa-
tion training
add to usual
care, 4 weeks,
not reported
Chaoul, 2018 YG 74 YG 49.5 ± 9.8 %White I–III Chemotherapy Yoga class (Tibetan CG1: stretch, BFI None
[24] CG1 68 CG1 50.4 ± 10.3 YG 58.1% yoga) + booster unclear, not
CG2 85 CG2 49 ± 10.1 CG1 68.7% class + home-based practice, report CG2:
Female CG2 65.4% 75–90 min + unclear + unclear, usual care,
unclear, unclear unclear, not
reported
Jong, 2018 [25] YG 47 YG 51 ± 8.0 Not reported I–III Chemotherapy/ Yoga class (Dru yoga) + home- Standard care, MFI/FQL/ None
CG 36 CG 51 ± 7.3 hormone based breathing and relaxation 12 weeks, EORTC-QLQ-
Female therapy/radio- exercise, 75 min + a minimum 13.8% were C30
therapy of 5 min, 1 time/week + daily, contaminated
12 weeks

BFI brief fatigue inventory, FSS fatigue severity scale, CFS cancer fatigue scale, CFS-D the German version of the cancer fatigue scale, CG control group, EORTC​-QLQ-C30/EORTC​-QOL-C30
European organization for research and treatment of cancer quality of life C30, FACIT-Fatigue the functional assessment of chronic illness therapy-fatigue, FACT​-Fatigue functional assessment
of cancer therapy-fatigue, FQL fatigue quality list, FSI fatigue symptom inventory, MFI multidimensional fatigue inventory, MFSI-SF multidimensional fatigue symptom inventory-short form,
VAS visual analog scale, YG yoga group
Breast Cancer Research and Treatment
Breast Cancer Research and Treatment

68.88 years. Ethnicity was reported in six studies; one study [13, 14] while one study had a high risk [12]. Four studies
was conducted with Caucasian populations; and four studies had a high risk of attrition bias [9, 14, 24, 25].
were performed in white populations. In the fifteen studies
that reported cancer stage, most were Stages I–III. Only four Analysis of overall effects
studies focused on breast cancer patients who had completed
anti-cancer treatment (post-treatment). The effect sizes with scores of 0.2–0.5, 0.5–0.8, and > 0.8
were considered small, medium, and large effects, respec-
Interventions tively [26].
The meta-analysis of the changes in fatigue scores from
Yoga types included Hatha yoga, Integrated yoga, Restora- the 17 studies indicated that yoga had a small but statisti-
tive yoga, Iyengar yoga, Viniyoga, Anusara yoga, Saty- cally significant beneficial effect, suggesting that yoga could
ananda yoga, Dru yoga, and Tibetan yoga. Forms of inter- mitigate fatigue in breast cancer patients to some extent
vention included supervised yoga class, home-based yoga, [SMD = − 0.31, 95% CI (− 0.52, − 0.10), P = 0.003] (Fig. 4).
home-based breathing and relaxation exercise, self-practice The funnel plot (Fig. 5) indicates that the publication bias
in the hospital, and centralized intervention for inpatients/ is mild, and the sensitivity analysis reveals that the model
decentralized intervention for outpatients. The supervised is relatively stable.
yoga class ranged from 40 to 90 min and took place 1–5
times per week for 4–24 weeks. Subgroup and sensitivity analysis

Status of treatment
Controls
The meta-analyses of the four studies with only post-treat-
Four studies used exercise as a control intervention [18–21]. ment patients [SMD = − 0.80, 95% CI (− 1.52, − 0.09),
One study used supportive counseling as a control interven- P = 0.03] and of the 13 studies with patients undergoing anti-
tion [9]. Health education classes were also used as the con- cancer treatment (intra-treatment) [SMD = − 0.25, 95% CI
trol intervention in one study [13]. The remaining 12 studies (− 0.47, − 0.03), P = 0.03] (Table 2) demonstrated that yoga
used standard care or usual care or non-intervention as the had a large and a small effect on fatigue in post-treatment
control intervention [8, 10–12, 14–17, 22–25]; in addition, and intra-treatment breast cancer patients, respectively.
more than one control group had been established in two of
these 12 studies [23, 24]. Type of yoga

Risk of bias in individual trial Twelve studies reported the yoga types, two of which were
Hatha yoga, three of which were Iyengar yoga, and the rest
The overall risk of bias as shown in Fig. 2 is moderate; were Integrated yoga, Restorative yoga, Viniyoga, Anusara
blinding of participants and personnel was not applicable for yoga, Tibetan yoga, Dru yoga, and Satyanada yoga, respec-
yoga intervention, so the risk of performance bias is high in tively. Therefore, the subgroup analyses of Hatha Yoga and
all studies. The individual risk of bias for each study is pre- Iyengar yoga were carried out. Hatha yoga had no effect
sented in Fig. 3. Ten studies reported the random sequence on fatigue in patients with breast cancer [SMD = 0.35, 95%
generation, one of which are high risk [19]. Seven studies CI (− 0.13, 0.83), P = 0.15], and Iyengar yoga also dem-
had a low risk of allocation concealment [9, 13, 14, 16–18, onstrated no effect [SMD = − 0.17, 95% CI (− 0.55, 0.21),
25]. In terms of detection bias, two studies had a low risk P = 0.37]. When we excluded Lötzke’s [20] study from the

Fig. 2  Overall risk of bias

assessment using the Cochrane
tool

13
 Breast Cancer Research and Treatment

Iyengar yoga subgroup because it had a high dropout rate

and used home-based yoga, and included intra-treatment
patients while the other two studies were conducted with
post-treatment patients [11, 13], the results indicated a large
effect [SMD = − 0.98, 95% CI (− 1.46, − 0.49), P < 0.0001].

Form of intervention

Two forms of intervention were analyzed, supervised yoga

class and supervised yoga class plus home-based yoga, the
former demonstrated a large beneficial effect [SMD = − 0.92,
95% CI (− 1.53, − 0.32), P = 0.003] while the latter had no
effect [SMD = 0.14, 95% CI (0.02, 0.25), P = 0.02] on CRF
in breast cancer patients.

Duration of yoga class

60-min and 90-min yoga class indicated a significant posi-

tive effect on CRF in breast cancer patients [SMD = − 1.20,
95% CI (− 2.28, − 0.12), P = 0.03]/[SMD = − 0.98, 95% CI
(− 1.46, − 0.49), P < 0.0001].

Length of program

Four-week and 12-week yoga class had no impact on

fatigue in patients with breast cancer [SMD = − 0.81,
95% CI (− 1.86, 0.24), P = 0.13]/[SMD = − 0.09, 95% CI
(− 0.30, 0.13), P = 0.43] (Table 2). The sensitivity analy-
sis indicated that 4-week yoga class had a moderate effect
[SMD = − 0.62, 95% CI (− 1.12, − 0.12), P = 0.02], i.e.,
the model is unstable; however, twelve-week yoga class
still had no effect [SMD = − 0.04, 95% CI (− 0.17, 0.09),
P = 0.51]. When we excluded Moadel’s [8], Cramer’s [17],
Lötzke’s [20], and Stan’s [21] studies from the 12-week
subgroup because they used home-based yoga, the outcome
indicated a small effect [SMD = − 0.45, 95% CI (− 0.75,
− 0.15), P = 0.003]. Six-week yoga class demonstrated a
moderate effect [SMD = − 0.68, 95% CI (− 1.07, − 0.29),
P = 0.0006] and eight-week yoga class manifested a large
effect [SMD = − 1.32, 95% CI (− 2.48, − 0.16), P = 0.03] on
fatigue.

Dimension of fatigue

Yoga has a significant beneficial effect on physical fatigue in

patients with breast cancer [SMD = − 0.83, 95% CI (− 1.34,
− 0.32), P = 0.001], a medium effect on cognitive fatigue
[SMD = − 0.63, 95% CI (− 0.90, − 0.35), P < 0.00001], and
a small effect on mental fatigue [SMD = − 0.47, 95% CI
(− 0.75, − 0.19), P = 0.001] (Table 2). However, yoga has no
impact on emotional fatigue [SMD = − 0.16, 95% CI (− 0.35,
Fig. 3  Risk of bias assessment by individual trials
0.03), P = 0.10], and the sensitivity analysis also indicated
no effect [SMD = − 0.16, 95% CI (− 0.35, 0.03), P = 0.09].

13
Breast Cancer Research and Treatment

Fig. 4  Overall effect of yoga on fatigue score changes in breast cancer patients

Adherence and contamination

Twelve studies reported the adherence to the yoga inter-

vention; adherence was higher than 80% in four studies
[11, 14, 16, 17]; five studies reported specific compliance
that could not be simply described in percentage terms [8,
9, 12, 24, 25], three of which had a low adherence [8, 12,
24]. Six studies did not report the adherence to yoga [15,
18–20, 22, 23]. Only one study reported the contamination
rate of the control group (13.8%) [25].

Adverse events

Eight studies reported the adverse events, five of which

had no adverse events [10, 12, 14, 24, 25] and three of
which reported a total of 26 adverse events [13, 17, 21].
Fourteen of these events may be related to the yoga inter-
Fig. 5  Funnel plot of comparison: overall effect of yoga on fatigue vention, but no serious events occurred.
score changes in breast cancer patients

13
 Breast Cancer Research and Treatment

Table 2  Subgroup and sensitivity analyses of yoga on fatigue score changes in breast cancer patients
Outcome type K ES Sample size Random-effects analysis Fixed-effects analysis
SMD 95% CI P SMD 95% CI P
YG CG L U L U

Status of treatment
Post-treatment 4 5 79 74 − 0.80 − 1.52 − 0.09 0.03 − 0.52 − 0.86 − 0.19 0.002
Intra- treatment 13 29 1033 997 − 0.25 − 0.47 − 0.03 0.03 − 0.16 − 0.25 − 0.07 0.0003
Type of yoga
Hatha yoga 2 3 122 86 0.35 − 0.13 0.83 0.15 0.24 − 0.04 0.52 0.1
Iyengar yoga 3 7 219 225 − 0.17 − 0.55 0.21 0.37 0.02 − 0.17 0.20 0.87
Form of intervention
Yoga class 7 11 214 216 − 0.92 − 1.53 − 0.32 0.003 − 0.77 − 0.98 − 0.57 < 0.00001
Yoga class + home- 6 14 579 593 0.14 0.02 0.25 0.02 0.14 0.02 0.25 0.02
based yoga
Duration of yoga class
60 min 2 5 128 132 − 1.20 − 2.28 − 0.12 0.03 − 0.94 − 1.21 − 0.66 < 0.00001
90 min 2 3 39 37 − 0.98 − 1.46 − 0.49 < 0.0001 − 0.98 − 1.46 − 0.49 < 0.0001
Length of program
4 weeks 2 2 34 33 − 0.81 − 1.86 0.24 0.13 − 0.62 − 1.12 − 0.12 0.02
6 weeks 2 3 80 75 − 0.68 − 1.07 − 0.29 0.0006 − 0.68 − 1.01 − 0.35 < 0.0001
8 weeks 3 5 88 88 − 1.32 − 2.48 − 0.16 0.03 − 1.41 − 1.77 − 1.05 < 0.00001
12 weeks 6 15 531 449 − 0.09 − 0.30 0.13 0.43 − 0.04 − 0.17 0.09 0.51
Dimension of fatigue
Physical fatigue 4 8 295 272 − 0.83 − 1.34 − 0.32 0.001 − 0.91 − 1.08 − 0.73 < 0.00001
Emotional fatigue 3 6 216 216 − 0.16 − 0.35 0.03 0.10 − 0.16 − 0.35 0.03 0.09
Cognitive fatigue 2 4 180 184 − 0.63 − 0.90 − 0.35 < 0.00001 − 0.63 − 0.84 − 0.41 < 0.00001
Mental fatigue 2 4 115 88 − 0.47 − 0.75 − 0.19 0.001 − 0.47 − 0.75 − 0.19 0.001

K number of studies, ES number of effect size, YG yoga group, CG control group, SMD standardized mean difference effect size, L lower, U
upper, Yoga class supervised yoga class, excluding studies that encouraged home-based yoga or exercise, self-practice in the hospital and used
intensive intervention/decentralized intervention

Discussion found that the results of subgroup analyses involving the

studies implementing home-based yoga were negative,
The meta-analysis including 2183 participants demonstrated such as the subgroup analyses of Hatha yoga, Iyengar yoga,
that yoga had a slight but statistically significant effect on and 12-week intervention; however, when we removed
mitigating fatigue in breast cancer patients. A former meta- these studies, the results became valid. Six studies adopted
analysis also indicated that yoga had a beneficial impact on supervised yoga class plus home-based yoga as the form of
fatigue in breast cancer patients [27], but it did not conduct intervention [8, 12, 14, 17, 20, 24] and one study used home-
subgroup analysis according to status of treatment, fre- based yoga only [21]. These seven studies contained certain
quency and duration of intervention, type of yoga, length of factors that might affect the analysis results. Moadel’s study
program, and dimension of fatigue. had a poor adherence-32% of the patients did not partici-
The included studies did not provide estimates of the pant in the intervention at all, and the low adherence was
effect size that could be pooled for subgroups of age, eth- associated with increased fatigue, radiotherapy, younger age
nicity, or cancer stage. Therefore, no subgroup analysis and no antiestrogen therapy [8]. Littman’s study also had a
was made according to these stratification factors. The fre- low adherence; patients attended an average of 19.6 facility-
quency and duration of home-based yoga and the frequency based classes (range 1–61; median: 20.5) and practiced at
of supervised yoga class could not be integrated, so we did home an average of 55.8 times (range 2–102; median: 62)
not conduct a subgroup analysis according to these factors. [12]. Loudon’s study had a small sample size (interven-
The subgroup analysis and the sensitivity analysis indi- tion group: 12, control group: 11) [14].The sample size of
cated that the supervised yoga class plus home-based yoga Cramer’s study was also small (intervention group: 19, con-
had no effect on fatigue in breast cancer patients; we also trol group: 21) [17]. Lötzke’s study had a high dropout rate

13
Breast Cancer Research and Treatment

and the missing data were imputed using Rubin’s multiple have significant impacts on study outcomes. Therefore,
imputation MI method; there were 92 participants in this further studies with more scientific methodology, larger
study; at t1, 2 participants dropped out from yoga group sample sizes, and a multi-center design are needed.
and 1 participant dropped out from control group; at t2, 29
and 25 patients dropped out from yoga group and control
group, respectively; for breast cancer patients, additional
interventions during cancer treatment may be too exhaust- Conclusions
ing; the high dropout rate may reflect a low acceptance
for supportive interventions during cancer treatment [20]. Yoga can be considered as an alternative therapy to relieve
Chaoul’s study indicated that the patient-instructor encoun- fatigue in breast cancer patients who have completed treat-
ter time was limited, which might account for the absence ment or are undergoing anti-cancer treatment. From this
of significant effects when comparing yoga to usual care; meta-analysis, we find that supervised yoga class can
the study’s booster sessions had a low adherence-51% of the reduce fatigue in breast cancer patients remarkably, a
participants did not attend any booster sessions [24]. Stan’s 6-week program had a medium effect on CRF; 60/90-min
study also demonstrated a poor adherence to yoga, only 39%; yoga class and a 8-week program had a significant effect;
the participants relied on self-discipline and the bi-weekly yoga could markedly reduce the physical fatigue in breast
phone calls as their source of motivation; home-based yoga cancer patients, had a medium impact on cognitive fatigue,
lacked scheduled sessions or the peer pressure, which might and a small effect on mental fatigue. In addition, future
lead to less motivation; the study also had a small sample researchers should address the issue of adherence and
size (intervention group: 18, control group: 16) [21]. Due to strengthen the supervision of yoga intervention to ensure
the above-mentioned factors, the results of analysis should its effect on CRF.
be treated with caution. As to what led to the negative results
in subgroup analysis, the above-mentioned factors or the
home-based yoga, we cannot easily conclude yet. Also, we Funding This study was supported by the Humanity and Social Sci-
ence Youth Foundation of Ministry of Education of China (Project
need to implement home-based yoga with caution, taking No. 18YJCZH164).
into account the safety problem, although no serious events
have been reported up to present. Compliance with ethical standards
Drawing upon the results of this meta-analysis, the clini-
cal professionals can formulate individualized yoga prac- Conflict of interest The authors declare that they have no conflicts of
tice program for breast cancer patients based on patients’ interest.
individual condition, e.g., supervised yoga class, 60/90 min/
Ethical approval All procedures performed in studies involving human
session with a duration of 6/8 weeks. However, more high- participants were in accordance with the ethical standards of the insti-
quality studies with high adherence, low dropout rate and tutional and/or national research committee and with the 1964 Helsinki
large sample size are still needed to confirm the best yoga declaration and its later amendments or comparable ethical standards.
type, the most valid form of intervention, the most reason-
able frequency and duration of yoga practice, and the most
suitable length of program in order to relieve CRF in breast
cancer patients. Appendix. A detailed search strategy
for Medline
Limitations of the current study
#1 “Breast Neoplasms” [Mesh] OR breast tumor [Title/Abstract]
Despite our comprehensive review of the literature on OR Breast Cancer [Title/Abstract] OR Breast Carcinoma [Title/
fatigue in breast cancer patients, the present study still Abstract] OR Mammary Cancer [Title/Abstract] OR Mammary
Carcinoma [Title/Abstract] OR Mammary Neoplasm [Title/
has some limitations. First, among the 17 studies included
Abstract] OR Mammary tumor [Title/Abstract]
in the meta-analysis, nine studies had a sample size of
less than 30 subjects. Second, the quality of the stud- #2 “Fatigue” [Mesh] OR Asthenia [Title/Abstract] OR Lassitude
ies included in this meta-analysis is mediocre, which [Title/Abstract]
may influence the results. Third, it is known that vari- #3 “Yoga” [Mesh] OR Yogic [Title/Abstract] OR Asana [Title/
Abstract]
ous symptoms, such as pain, sleep disturbance, emotional
#4 randomized controlled trial [Publication Type]
distress, nutrition, and level of activity among others, can
#5 #1 AND #2 AND #3 AND #4
impact CRF level; however, several included studies did
not report information regarding these factors, which may

13
 Breast Cancer Research and Treatment

References 16. Taso Chao-Jung, Lin Huey-Shyan, Lin Wen-Li et al (2014) The
effect of yoga exercise on improving depression, anxiety, and
fatigue in women with breast cancer: a randomized controlled
1. Ghoncheh Mahshid, Pournamda Zahra, Salehiniya Hamid (2016)
trial. J Nurs Res 22(3):155–163
Incidence and mortality and epidemiology of breast cancer in the
17. Cramer Holger, Rabsilber Sybille, Lauche Romy et al (2015)
world. Asian Pac J Cancer Prev 17(S3):43–46
Yoga and meditation for menopausal symptoms in breast cancer
2. Ziegler RG et al (2008) Increasing breast cancer incidence in
survivors-a randomized controlled trial. Cancer 121:2175–2184
China: the numbers add up. J Natl Cancer Inst 100(19):1339–1341
18. Yağlı Naciye Vardar, Şener Gül, Arıkan Hülya et al (2015) Do
3. Berger Ann M, Mooney Kathi, Alvarez-Perez Amy et al (2015)
yoga and aerobic exercise training have impact on functional
Cancer-related fatigue, version 2.2015: clinical practice guidelines
capacity, fatigue, peripheral muscle strength, and quality of life
in oncology. J Natl Compr Canc Netw 13(8):1012–1039
in breast cancer survivors? Integr Cancer Ther 14(2):125–132
4. Bardwell WA, Ancoli-Israel S (2008) Breast cancer and fatigue.
19. Yagli Naciye Vardar, Ulger Ozlem (2015) The effects of yoga on
Sleep Med Clin 3:61–71
the quality of life and depression in elderly breast cancer patients.
5. Abrahams HJG, Gielissen MFM, Schmits IC et al (2016) Risk fac-
Complement Ther Clin Pract 21(2015):7–10
tors, prevalence, and course of severe fatigue after breast cancer
20. Lötzke Désirée, Wiedemann Florian, Recchia Daniela Rodrigues
treatment: a meta-analysis involving 12327 breast cancer survi-
et al (2016) Iyengar-yoga compared to exercise as a therapeutic
vors. Ann Oncol 27:965–974
intervention during (neo)adjuvant therapy in women with stage
6. Smith KB, Pukall CF (2009) An evidence-based review of yoga as
I-III breast cancer: health-related quality of life, mindfulness, spir-
a complementary intervention for patients with cancer. Psychoon-
ituality, life satisfaction, and cancer-related fatigue. Evid-Based
cology 18(5):465–475
Complement Alternat Med 2016:1–8
7. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group
21. Stan Daniela L et al (2016) Randomized pilot trial of yoga versus
(2009) Preferred reporting items for systematic reviews and meta-
strengthening exercises in breast cancer survivors with cancer-
analyses: the PRISMA statement. Ann Intern Med 151:264–269
related fatigue. Support Care Cancer 24:4005–4015
8. Moadel Alyson B, Shah Chirag, Wylie-Rosett Judith et al (2007)
22. Jin Cuifeng, Wang Lily, Wang Bei (2017) Effect of yoga exercise
Randomized controlled trial of yoga among a multiethnic sample
on cancer-related fatigue and quality of life in patients with breast
of breast cancer patients: effects on quality of life. J Clin Oncol
cancer during chemotherapy. Nurs Integr Tradit Chin West Med
25(28):4387–4395
3(4):12–15
9. Hosakote Vadiraja S, Raghavendra Rao M, Hongasandra Nagendra
23. Zeng Jinfang (2017) Effect of yoga combined with music relaxa-
R et al (2009) Effects of yoga on symptom management in
tion training on cancer-related fatigue in patients with breast can-
breast cancer patients: a randomized controlled trial. Int J Yoga
cer chemotherapy. J Clin Nurs Pract 2(19):1–2
2(2):73–79
24. Chaoul Alejandro, Milbury Kathrin, Spelman Amy et al (2018)
10. Danhauer Suzanne C, Mihalko Shannon L, Russell Gregory B et al
Randomized trial of tibetan yoga in breast cancer patients under-
(2009) Restorative yoga for women with breast cancer: findings
going chemotherapy. Cancer 124(1):36–45
from a randomized pilot study. Psychooncology 18(4):360–368
25. Jong Miek C, Boers Inge, Arjan P, van der Velden Schouten et al
11. Banasik Jacquelyn et al (2011) Effect of Iyengar yoga practice on
(2018) A randomized study of yoga for fatigue and quality of life
fatigue and diurnal salivary cortisol concentration in breast cancer
in women with breast cancer undergoing (neo) adjuvant chemo-
survivors. J Am Acad Nurse Pract 23:135–142
therapy. J Alternat Complement Med 24(9):942–953
12. Littman Alyson J, Bertram Lisa Cadmus, Ceballos Rachel et al
26. Conn VS, Hafdahl AR, Porock DC, McDaniel R, Nielsen PJ
(2012) Randomized controlled pilot trial of yoga in overweight
(2006) A meta-analysis of exercise interventions among people
and obese breast Cancer survivors: effects on quality of life and
treated for cancer. Support Care Cancer 14(7):699–712
anthropometric measures. Support Care Cancer 20(2):267–277
27. Zhang Qi et al (2015) Meta-analysis of the effect of yoga on
13. Bower Julienne E et al (2012) Yoga for persistent fatigue in
cancer-related fatigue in breast cancer patients. Chin J Mod Nur
breast cancer survivors: a randomized controlled trial. Cancer
21(28):3380–3386
118(15):3766–3775
14. Loudon Annette, Barnett Tony, Piller Neil et al (2014) Yoga man-
Publisher’s Note Springer Nature remains neutral with regard to
agement of breast cancer-related lymphoedema: a randomised
jurisdictional claims in published maps and institutional affiliations.
controlled pilot-trial. Complement Alternat Med 14:214
15. Wang G et al (2014) Effect of yoga on cancer related fatigue in
breast cancer patients with chemotherapy. Zhong Nan da Xue Xue
Bao 39(10):1077–1082

13