R

Msr

Ka

b

c

C

12

3

4

5

6

789

h1

Critical Reviews in Oncology/Hematology 111 (2017) 66–80

Contents lists available at ScienceDirect

Critical Reviews in Oncology/Hematology

jo u r n al homep age: www.elsev ier .com/ locate /c r i t revonc

eview

anagement of aromatase inhibitor induced musculoskeletalymptoms in postmenopausal early Breast cancer: A systematiceview and meta-analysis

ate Robertsa,∗, Kirsty Rickettb, Ristan Greerb,c, Natasha Woodwarda,b,c

Department of Medical Oncology, Mater Health Services, Brisbane, AustraliaUniversity of Queensland, Brisbane, AustraliaMater Medical Research Institute Mater Health Services, Brisbane, Australia

ontents

. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672.1. Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672.2. Study selection: inclusion and exclusion criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

2.2.1. Type of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672.3. Types of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682.4. Types of intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682.5. Types of outcome measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682.6. Data synthesis and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682.7. Methodological quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.1. Search results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.2. Methodological quality of selected studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.3. Overall characteristics of selected studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.4. Analysis of acupuncture interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68

3.4.1. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.1. Analysis of pharmacological interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714.1.1. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725.1. Analysis of complementary interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5.1.1. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

6.1. Analysis of physical therapy interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736.1.1. Characteristics of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74Appendix A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75

PubMed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75EMBASE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76CENTRAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77CINAHL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

∗ Corresponding author at: Department Medical Oncology, Mater Health Services, South Brisbane, 4101, Australia.E-mail address: Kate.Roberts@health.qld.gov.au (K. Roberts).

ttp://dx.doi.org/10.1016/j.critrevonc.2017.01.010040-8428/© 2017 Elsevier B.V. All rights reserved.

a

ARR3A

KALEAAMAB

1

tpoCh(iibwatmtrt2mtre

la2pdet2Ttfoyseepced

ha

K. Roberts et al. / Critical Reviews in Oncology/Hematology 111 (2017) 66–80 67

r t i c l e i n f o

rticle history:eceived 4 October 2016eceived in revised form0 November 2016ccepted 18 January 2017

eywords:romatase inhibitorsetrozolexemestanenastrozoleIMSSusculoskeletal symptoms

rthralgiasreast cancer

a b s t r a c t

Aromatase Inhibitors (AI) are widely used for the adjuvant treatment of hormone receptor positive breastcancers in the post-menopausal population. AI are often associated with significant joint and muscularsymptoms; symptoms that are commonly referred to as aromatase inhibitor-associated musculoskeletalsyndrome (AIMSS). AIMSS adversely impacts health-related quality of life of many patients, and reduces AIcompliance. Although there are informal practice recommendations, the limited current level of evidencefor management of AIMSS for breast cancer patients on aromatase inhibitors has made development offormal guidelines challenging, and remains an unmet need. This is the first systematic review to considerthe evidence for all pharmacological and non-pharmacological interventions in the treatment of AIMSS,including physical therapy, acupuncture and complementary therapies.

© 2017 Elsevier B.V. All rights reserved.

. Introduction

Aromatase Inhibitors (AI) are recommended for the adjuvantreatment of hormone receptor positive breast cancers in theost-menopausal population. These agents block the synthesis ofestrogen by inhibition of peripheral aromatase (Miller et al., 2003).ompared with Tamoxifen, third generation aromatase inhibitorsave been shown to significantly improve disease free survivalDFS) (Cuzick et al., 2010; Regan et al., 2011; Ingle et al., 2006), andnclude the steroidal inhibitor exemestane, and the nonsteroidalnhibitors, anastrozole and letrozole. In the 2013 meta-analysisy Aydiner et al. (Aydiner, 2013), five years of adjuvant therapyith aromatase inhibitors improved DFS (HR 0.89, p = 0.001), and

lso overall survival (OS) (HR 0.92, p = 0.046) when compared toamoxifen. Aromatase inhibitors have also demonstrated improve-

ent in DFS, OS and distant metastasis rate when sequenced withamoxifen (HR 0.70, p < 0.001; HR 0.81, p = 0.003, HR 0.74, p < 0.001espectively), and an improvement in DFS as extended adjuvantreatment after 5 years of tamoxifen (HR 0.62, p = 0.001) (Aydiner,013). Recent evidence has revealed a benefit of continuing aro-atase inhibitors for a period of 10 years, as reported in the MA.17R

rial, which displayed significant improvement in breast cancerecurrence rates, and decreased contralateral breast cancer (Gosst al., 2016).

Aromatase inhibitors are associated with joint and muscu-ar symptoms, commonly referred to as aromatase inhibitor-ssociated musculoskeletal syndrome (AIMSS) (Lintermans et al.,013). AIMSS adversely impacts on the quality of life of manyatients. Studies recently investigating AIMSS have shown inci-ence of musculoskeletal symptoms to be as much as 50% (Colemant al., 2008; Laroche et al., 2014; Menas et al., 2012), higher thanhe pivotal aromatase inhibitor trials with rates of approximately0–35% (Muss et al., 2008; Coates et al., 2007; Howell et al., 2005).he prevalence of musculoskeletal symptoms impacts the long-erm care of these patients. Analysis of longitudinal claims datarom three American commercial health programs revealed sub-ptimal adherence to anastrozole in 19–28% of patients in their firstear of treatment (Partridge et al., 2008). These statistics are con-istent with other studies of aromatase inhibitor adherence (Hadjit al., 2014; Hershman et al., 2011; Presant et al., 2007; Henryt al., 2012), which report a significant percentage of patients dis-laying early discontinuation of treatment. There are importantlinical implications of this data, as non-compliance with adjuvantndocrine therapies in early breast cancer has been shown to be

(Burstein, 2007). There may be additional extra-articular symp-toms present, such as myalgia, fibromyalgia, neuropathy and carpaltunnel syndrome (Sestak et al., 2009). MRI studies conducted onpatients taking aromatase inhibitors have shown the develop-ment of tenosynovial changes and increased intra-articular fluidin patients with AIMSS (Lintermans et al., 2013). Most of thesymptoms will develop within the first two to three months ofAI treatment (Burstein, 2007; Mao et al., 2009a). This systematicreview aims to summarise the recent literature on the symp-tom management intervention strategies for AIMSS. Meta-analyseshave been conducted where feasible.

2. Methods

2.1. Search strategy

A systematic search of the electronic literature was designedand conducted by an information specialist (KR) to identify the rel-evant evidence. The following databases were searched: PubMed,EMBASE, CINAHL and CENTRAL. Controlled terminology (MESH,EMTREE, CINAHL headings) and free text words were used. Googlescholar was also searched for unpublished literature. The finalsearch of all the databases was conducted on 24th February 2016.Reference lists of relevant review articles and of the full textreviewed papers were also cross checked and any relevant papersincluded for review. The complete search strategies for all thedatabases can be found in Appendix A.

2.2. Study selection: inclusion and exclusion criteria

2.2.1. Type of studiesAlthough the best type of study to assess the efficacy of an inter-

vention is a randomised controlled trial (RCT), the scope of studiesfor inclusion in this review has been expanded. This is to reflectthe recognition that there are very few RCT in the area, and to beinclusive of as many intervention types as possible to inform clini-cal practice and respond to patient enquiries. Therefore, all clinicaltrials (prospective and retrospective), cohort and case control stud-ies and preventative trials were considered. Conference abstractswere included, but where a later full paper has been published,the abstract was excluded and replaced with the full paper. Let-ters to the editor detailing clinical trial results were also included.

etrimental to the patients’ survival (Hershman et al., 2011).AIMSS usually presents as symmetrical pain or soreness in the

ands, knees, hips, lower back, shoulders, and/or feet. It is oftenssociated with early-morning stiffness and difficulty sleeping

Conference abstracts and letters to the editor were only consid-ered in the narrative analysis and were not included in the risk ofbias assessment or meta-analysis as there was not enough infor-

6 Oncol

mssaw

2

mimwPdpi

2

inM

2

ifpdts

2

abHhhtmst

2

aww7m

3

3

dFc

8 K. Roberts et al. / Critical Reviews in

ation to make an accurate analysis. Case studies and small caseeries were excluded. Papers detailing protocols only, as well asystematic reviews were excluded from the review, although thesere considered in the discussion. Only papers published in Englishere considered.

.3. Types of participants

Women with stage I–III Breast Cancer on an adjuvant treat-ent with any aromatase inhibitor with, or at risk of, AIMSS were

ncluded. AIMSS was defined as any new onset, or worsening, of anyusculoskeletal symptom after commencement of an AI. Womenith advanced/metastatic breast cancer (Stage IV) were excluded.

apers which did not clearly define the use of aromatase inhibitorsistinct to other hormonal therapies were excluded, along withapers which included endocrine therapies other than aromatase

nhibitors

.4. Types of intervention

All types of symptom management interventions for AIMSSn this population were considered including – pharmacological,on-pharmacological and CAM (Complementary and Alternativeedicine).

.5. Types of outcome measures

Primary outcomes and secondary outcomes included themprovement in AIMSS (pain, stiffness, mobility or functionality)rom baseline, the improvement in persistence and compliance ofatients continuing to take their aromatase inhibitor medicationue to the intervention, the reduction in incidence of AIMSS, andhe adverse events in relation to the intervention treating AIMSSymptoms.

.6. Data synthesis and analysis

Where sufficient quantitative results were reported, meta-nalysis was performed. I2 was used to measure heterogeneityetween studies, as per the Cochrane handbook (Cochraneandbook, 2011). An I2 value of 50–75% is defined as substantialeterogeneity and an I2 value of ≥75% is defined as considerableeterogeneity. Heterogeneity was expected between studies, andherefore a random-effects meta-analysis model was used for the

eta-analyses. A separate meta-analysis was attempted for eachub-group of intervention. R programming software was used forhe statistical analysis (Schwarzer , 2016).

.7. Methodological quality

RCTs were assessed using the Jadad Scale (Olivo et al., 2008). Tri-ls were deemed high quality studies if score 3–5, whilst score 0–2as deemed low quality (Jadad et al., 1996). Case control studiesere assessed using the Newcastle-Ottawa Scale, where a score of

–9 indicates high methodological quality, a score of 4–6 indicatesoderate quality and a score of 0–3 indicates low quality.

. Results

.1. Search results

The search retrieved 1389 articles and after the removal of 458uplicates, 931 remaining abstracts were screened, as shown inig. 1. After 836 of these abstracts were excluded, 95 full text arti-les were assessed, with 38 meeting the inclusion criteria. 57 papers

ogy/Hematology 111 (2017) 66–80

were excluded: In 17 studies the relevant outcomes were not cov-ered; in 12 studies hormonal therapy was not distinguished asAI; in 11 studies the abstracts were superseded by later full textpapers; 10 papers were the wrong study design or publication type;6 papers had the wrong patient setting; and 1 paper was not inEnglish.

3.2. Methodological quality of selected studies

Studies were unable to be assessed if they had only been pub-lished as an abstract, due to lack of available information. Outof 17 RCTs, only 11 could be adequately assessed for method-ological quality, as the rest were only published in abstract form.Of the 11 assessable RCTs, eight studies had a high Jadad score≥3 (Shapiro et al., 2016; Rastelli et al., 2011; Oh et al., 2013;Mao et al., 2014; Hershman et al., 2015; Crew et al., 2010; Baoet al., 2013; Fields, 2015). Three studies scored poorly on theJadad scale for methodological quality (Crew et al., 2007; Irwinet al., 2015; Liu et al., 2014). Only six cohort studies had full-text available for assessment with the Newcastle Ottowa Scale.Of these six studies, two were assessed as high methodologicalquality (Prieto-Alhambra et al., 2011; Cantarero-Villanueva et al.,2013); two studies were median methodological quality (Presantet al., 2007; Khan et al., 2010; Muslimani et al., 2009); and onestudy was of poor methodological quality (Xepapadakis et al.,2010).

3.3. Overall characteristics of selected studies

38 studies were included in the final analysis (see Table 1). Theseincluded 18 randomised control trials (RCTs)/controlled clinical tri-als (CCT), 14 pre/post studies, and 6 cohort studies. Studies werepublished between 2007 and 2016. The countries in which the stud-ies were conducted include: United States (n = 26), Japan (n = 2),Spain (n = 2), China (n = 2), Canada (n = 1), England (n = 1), Australia(n = 1), Greece (n = 1), France (n = 1) and Italy (n = 1). Of the trialswhich reported median ages of participants, the median age was59.5 years (range 29–89). The scoring systems used across trialswere extremely diverse, including Health Assessment Question-naire (HAQ), Visual Analogue Scale (VAS), Brief Pain Inventory (BPI),Western Ontario and McMaster Universities Osteoarthritis Index(WOMAC), Modified Score for the assessment of Chronic Rheuma-toid Affections of the Hands (M-SACRAH), Functional Assessmentof Cancer Therapy – General (FACT-G), Arthritis Impact Measure-ment Scale (AIMS2), Medical Outcome Study Short Form 36 (SF36), Pain self-efficacy questionnaire (PSEQ), OMERACT-OARSI crite-ria, Disabilities of the Arm, Shoulder and Hand (DASH), AUStralianCANadian Osteoarthritis Hand Index (AUSCAN), Breast Cancer Pre-vention Trial –; Musculoskeletal Symptom (BCPT-MS), 5 pointLikert Scale, and the use of an electronic algometer and hand gripstrength.

The studies were analysed in four separate subgroups: Com-plementary Therapies; Acupuncture; Physical Therapies andPharmacological Interventions.

3.4. Analysis of acupuncture interventions

3.4.1. Study characteristicsSix studies were included that investigated the use of acupunc-

ture for AIMSS, involving 221 patients in total. There were five RCTs,three of which investigated the use of acupuncture (Crew et al.,2010; Bao et al., 2013; Crew et al., 2007) and two of which inves-

tigated the use of electro-acupuncture (Oh et al., 2013; Mao et al.,2014). There was also one single-arm pilot study, which investi-gated the use of electroacupuncture (Mao et al., 2009b). Of the RCTs,three studies investigated true acupuncture/electro-acupuncture

K. Roberts et al. / Critical Reviews in Oncology/Hematology 111 (2017) 66–80 69

Table 1All included studies in the analysis.

Intervention &Author

Type n Arms Significant Outcome

AcupunctureAcupuncture

Bao et al. (2013)RCT 51 Real vs Sham (8 weeks) No. Similar VAS scores at 8 weeks

between real and sham arms (p 0.31)Acupuncture

Crew et al. (2007)RCT 21 Real (RA) vs Sham (SA)

6 weeks then crossoverYes, RA better at 6 weeks (values notgiven). Not sustained 12 weeks

AcupunctureCrew et al. (2010)

RCT 38 Real vs Sham (6 weeks) Yes. Worst pain 3 vs 5.5(p < 0.001) at 6 weeks

ElectroacupunctureOh et al. (2013)

RCT 32 Real vs Sham (6 weeks) No. No difference in pain betweengroups (no values given)

ElectroacupunctureMao et al. (2014)

RCT 67 Real vs Sham vs waitlist (8 weeks) Yes. Decrease in pain severity in EA(−2.2) vs WLC (−0.2, p = 0.0004), week8. Benefits continue week 12.

ElectroacupunctureMao et al.(2009b)

Pre/post 12 6 weeks Yes. Decrease in pain severity (5.3–1.9;p < 0.001)

PharmacologicalTestosterone

Birrell and Tilley(2010)

RCT 90 Placebo vs 40 mg vs 80 mg daily Yes. VAS score decreased 70% in 80 mgarm, compared to 35% decrease inplacebo arm (p = 0.04)

EtoricoxibRosati et al.(2011)

RCT 182 Etoricoxib 60 mg/day vs placebo Yes. Improved pain. 31% MSSsymptoms vs 76% (RR 2.1 p = 0.002)

CalcitoninLiu et al. (2014)

RCT 82 200IU/day salmon calcitonin daily Yes. Improvement in VAS scores inboth groups (placebo −1.00; calcitonin−3.00). Difference between groupsp < 0.01.

SwitchKadakia et al.(2016)

Pre/post 83 Exemestane or letrozole; crossoverif intolerant

62% continued second aromataseinhibitor at 6 months.

SwitchBriot et al. (2010)

Pre/post 179 Anastrozole switched to letrozole Yes. 19% decrease pain scores(p < 0.001)72% continued letrozole 6 months.

DuloxetineHenry et al.(2011)

Pre/post 29 30 mg 7 days, then 60 mg 21 days,then 60 mg bid

Yes. Mean decrease pain severity 60%(p < 0.001; 95% CI 49–73%)

PrednisoloneKubo et al. (2012)

pre/post 27 5 mg prednisolone for 1 week. Yes. 52% reported improved VAS at 2months post prednisolone use.No p-values given

ThymosinZhang et al.(2010)

Pre/post 16 Thymosin �1 1.6 mg twice weeklyfor 4 weeks

Yes. Decreased BPI worst pain scores(5.7–3.4, p < 0.001)

BisphosphonatesSanta-Maria et al.(2014)

Cohort 59 Zoledronic acid 4 mg baseline and6 months

Yes. 37% on zoledronic acid had AIMSS,compare to 61% in historical cohort(p < 0.001)

BisphosphonatesMuslimani et al.(2009)

RetrospectiveCohort

316 Calcium and Bisphosphonate vsnone

Yes. Associated between AIMSSsymptoms and low bone mineraldensity (p < 0.001)

Diuretic TherapyXepapadakiset al. (2010)

RetrospectiveCohort

288 Diuretic therapy vs no diuretics Yes. 7% on diuretics had arthralgia,compared with 16% not on diuretics(p0.01)

Analgesics/supplementsPresant et al.(2007)

RetrospectiveCohort

56 Chart review pain meds andeffectiveness

50% obtained relief from NSAIDs

Complementary TherapiesBlue Citrus Herbal

(BCH)Massimino et al.(2011)

RCT 31 BCH vs Placebo (P) then crossover90 days each

Yes. Both arms experienced decreasedVAS (p < 0.02). No p value given forBCH vs P

O3FAHershman et al.(2015)

RCT 249 O3FA vs placebo24 weeks

Yes. Improved BPI worst pain in O3FAarm (−2.23, p < 0.001) and placebo arm(−1.81, p < 0.001). No differencebetween arms (p = 0.52)

O3FALustberg et al.(2015)

RCT 44 O3FA vs placebo24 weeks

No difference in mean BPI-SF painscores. Less interference of pain inO3FA arm (−0.72, p = 0.08)

Vitamin DKhan et al. (2012)

RCT 160 30000IU D3wkly vs placebo for 24weeks

No. 51% in placebo arm vs 37% invitamin D arm had MS event (p = 0.069)

Vitamin DRastelli et al.(2011)

RCT 60 Vit D2 50000IU vs placebo weekly8–16 weeks then mthly for 4months

Benefit in BPI worst pain in Vit D armat 2 months (p = 0.0045).No difference between groups inregards to pain at 4 or 6 months.

70 K. Roberts et al. / Critical Reviews in Oncology/Hematology 111 (2017) 66–80

Table 1 (Continued)

Intervention &Author

Type n Arms Significant Outcome

Vitamin DShapiro et al.(2016)

RCT 113 600 IU Vit D3 vs 4000 IU Vit D3daily for 6 months

No. Change in BCPT-MS scores at 6months: −0.5 in 600IU D3 vs −0.2 in4000 IU D3, p = 0.38

Vitamin DKhan et al. (2010)

Cohort 51 Vit D3 50000IU weekly for 12weeks if 25OHD <40 ng/ml vsstandard D3/Calcium

Yes. Difference in pain between VitD <66 ng/ml (52%) vs >66 ng/ml (19%;p = 0.026).

Vitamin DPrieto-Alhambraet al. (2011)

Cohort 290 16000 IU Vit D3 q2wkly if25(0H)D <30 ng/ml vs Vit D3 800IU daily

Joint pain less likely if25(OH)D > 40 ng/ml (p < 0.008). VASscores increased in entire cohort(p < 0.001)

Vitamin EKiyomi et al.(2015)

Pre/post 62 Vit E (150 mg/day) for 30 days Yes. Mean osteoarthropathy scores(scoring 0–3) improved with Vitamin E(p = 0.0178)

Glucosamine/chondroitinGreenlee et al.(2013)

Pre/post 37 Glucosamine(1500 mg) + chondroitin (1200 mg)daily. 24 weeks

Yes. Benefit in BPI worst pain (−1.2,p = 0.02); M-SACRAH pain (−13.8,p = 0.0008) & WOMAC pain (−10.7,p = 0.02).

Physical TherapiesAerobic/Resistance

Irwin et al.(2015)

RCT 121 12/12 supervised vs usual care Yes. Decreased pain by 1.6 points vs 0.2points (p < 0.001)

Aerobic/ResistanceLohrisch et al.(2011)

RCT 20 48/52. Step-down supervision vsUC

No. No difference between groups. Nop values given for results.

Aerobic/ResistanceLash et al. (2011)

Pre/post 14 Supervised, 2–3 times/week, 8/52 Yes. Pain in multiple joints decreased(p < 0.05). Scores not given.

Aerobic/ResistanceDeNysschen et al.(2014)

Pre/post 26 8/52 home-based exercise Yes. Improvement in pain by AIMS2,−2.7 (p = 0.01).

AquaticCantarero-Villanueva et al.(2013)

CCT 40 2/12 hydrotherapy vs waitlist Yes. Improved pressure pain thresholdin treatment arm (p < 0.05). No benefitin waitlist arm.

Nordic WalkingFields, (2015)

RCT 40 6/52 supervised followed by 6/52self-managed walking vs waitlistcontrol

No. Pain scores changed −1.1 Nordicwalking vs −2.4 control group(p = 0.10)

WalkingNyrop et al.(2014)

Pre/post 20 Self-directed walking program,6/52

No. Mean joint pain decreased 10%(p0.63)

Tai ChiGalantino et al.,2013)

Pre/post 12 Supervised tai chi, 8/52 No. Difference in pain severity −1.04(p = 0.058)

YogaJacobsen et al.(2015)

Pre/post 10 Supervised yoga, 12/52 Yes. Improved BPI pain severity −1.35(p = 0.015)

YogaGalantino et al.(2012)

Pre/post 10 Supervised, 8/52 Yes. BPI pain severity reduced(3.90–2.79; p < 0.05)

VAS = visual analogue scale, EA = electroacupuncture, WLC = waitlist control, MSS = musculoskeletal symptoms, Vit = vitamin, BPI = brief pain inventory, AIMSS = Aromatasei atoryM AH = MW = Arth

veaea(bdsp2a1poi

nhibitor-induced musculoskeletal symptoms, NSAIDs = non-steroidal anti-inflammS = Breast Cancer Prevention Trial-Musculoskeletal Symptoms subscale, M-SACROMAC = Western Ontario and McMaster Universities Osteoarthritis Index, AIMS2

ersus sham acupuncture (Oh et al., 2013; Crew et al., 2010; Baot al., 2013); one study had three arms consisting of true electro-cupuncture, sham electro-acupuncture and a waitlist control (Maot al., 2014) and one trial was a crossover design, investigatingcupuncture versus observation, then crossover after six weeksCrew et al., 2007). The methods of sham acupuncture differedetween trials, with some trials using sham, non-penetrating nee-les on real acupuncture points (Oh et al., 2013), and others usingham, non-penetrating needles at non-acupuncture, non-triggeroints (Mao et al., 2014; Bao et al., 2013). Crew et al. (Crew et al.,010), used superficial needle insertion at non-acupuncture pointss the control. The median sample size for the studies was 37 (range2–67). The primary outcome of all the studies included change inain scores after the intervention. One study also listed primary

utcomes as perceived benefit of acupuncture, hand strength andnflammatory markers (ESR and CRP) (Oh et al., 2013).

drugs, O3FA = omega 3 fatty acids, BPI-SF = brief pain inventory, short-form, BCPT-odified Score for the assessment of Chronic Rheumatoid Affections of the Hands,ritis Impact Measurement Scale.

4. Results

In the two RCTs by Crew et al. (2010, 2007), there was a reportedbenefit in the symptoms of AIMSS with the use of acupuncture. Inthe 2007 trial of 21 patients (Crew et al., 2007), patients underwentsix weeks of acupuncture followed by six weeks of observation,or vice versa. The mean BPI worst pain score at baseline was 5.3compared with the mean BPI worst pain score after acupunctureof 3.3 (p = 0.008). The benefits of acupuncture did not persist aftersix weeks of observation. In the 2010 RCT by Crew et al. (2010), 43patients were randomised to either real or sham acupuncture for sixweeks. There was a difference in pain scores at six weeks betweentrue acupuncture and sham acupuncture arms, with mean BPI-SFworst pain scores 3.0 for true acupuncture versus 5.5 for sham

acupuncture (p = 0.002). Similar benefits were seen in pain severity(2.59v 4.53; p < 0.001) and pain-related interference (2.48v 4.54;p < 0.002). No follow-up was performed after acupuncture cessa-

K. Roberts et al. / Critical Reviews in Oncology/Hematology 111 (2017) 66–80 71

ma Fl

ttaIrw(vprsetaei

Fig. 1. Pris

ion. In all the remaining RCTs investigating the use of acupuncturehere was no statistical difference in pain outcomes between realnd sham arms (Oh et al., 2013; Mao et al., 2014; Bao et al., 2013).n the trial by Bao et al. (2013), 47 patients were randomised toeal or sham acupuncture for eight weeks. After eight weeks, thereas no difference between treatment arms in either HAZ-DI scores

p = 0.15) or VAS scores (p = 0.31). Oh et al. (2013), investigated realersus sham electroacupuncture for six weeks of treatment in 32atients. There was no difference between real and sham arms inegards to pain, function and stiffness using WOMAC scores, or paineverity and interference using BPI-SF scores. In the trial by Maot al. (2014), 67 patients were enrolled into a three arm RCT, inves-

igating real and sham electroacupuncture versus a waitlist controlrm, for eight weeks. Both true electroacupuncture (EA) and shamlectroacupuncture (SA) arms revealed a significant improvementn pain severity compared with the waitlist control arm (−2.0 vs

ow Chart.

−0.2, p = 0.0004), but there was no difference between EA and SAarms.

Two studies could be included in the method of meta-analysis, asthey used the same scoring systems within their studies (Mao et al.,2014; Crew et al., 2010) (Fig. 2). There was significant between-study heterogeneity for the effects of acupuncture on BPI-SF worstpain score (I2 = 79%). The overall mean difference in worst painscores after acupuncture, using the random effects model was−0.98 (95% CI, −;3.01–1.06).

4.1. Analysis of pharmacological interventions

4.1.1. Study characteristicsTwelve studies were included that investigated pharmacologi-

cal interventions for the management of AIMSS, including 3 RCTs,5 pre/post studies, and 4 cohort studies. Pharmacological ther-

72 K. Roberts et al. / Critical Reviews in Oncology/Hematology 111 (2017) 66–80

andar

aT22e22boiumnsotstabwsoa2e

5

wRoac(dTt(2iavthTptmd&

Fig. 2. Effect of acupuncture on AIMSS using BPI-SF. SD, st

pies used were diverse, and included testosterone (Birrell andilley, 2010), etoricoxib (Rosati et al., 2011), calcitonin (Liu et al.,014), duloxetine (Henry et al., 2011), prednisolone (Kubo et al.,012), thymosin (Zhang et al., 2010), bisphosphonates (Muslimanit al., 2009; Santa-Maria et al., 2014), diuretics (Xepapadakis et al.,010) and switching of aromatase inhibitor therapy (Kadakia et al.,016; Briot et al., 2010). There were 1407 patients analysed in totaletween all the pharmacological trials, with a median sample sizef 82.5 patients (range 16–316). The RCTs used a matched placebon the control arms, and one of the retrospective cohort studiessed controls from the ELPh Trial (Exemestane and Letrozole Phar-acogenetics Trial) (Santa-Maria et al., 2014). The other studies did

ot include a control arm. The primary outcome for majority of thetudies was either the impact of the pharmacological interventionn musculoskeletal symptoms, or the prevention of musculoskele-al symptoms through use of the pharmacological intervention. Thetudy by Liu et al., had a primary outcome of the efficacy of calci-onin as therapy for osteoporosis in patients with bone pain duringnastrozole therapy (Liu et al., 2014). In the study of switch therapyy Briot et al. (2010), the primary outcome was the percentage ofomen who discontinued letrozole secondary to musculoskeletal

ymptoms, after switching from anastrozole. Likewise, in the studyf switch therapy by Kadakia et al., the primary outcome was toler-nce of, and persistence with aromatase inhibitors (Kadakia et al.,016). In the study of etoricoxib (Rosati et al., 2011), the primarynd-point was 5 years event free survival.

. Results

Patients in all three RCTs experienced a reduction in pain, whichas the primary outcome (Liu et al., 2014; Birrell and Tilley, 2010;osati et al., 2011). In the trial by Birrell and Tilley (2010), which hasnly been published as an abstract, 80 mg testosterone resulted in

70% decrease in Visual Analogue Scores (VAS) scores at 3 months,ompared to only 35% decrease in VAS scores in the placebo armp = 0.04). The third arm in this trial, testing 40 mg testosterone,id not result in a substantial decrease in pain scores (p = 0.06).he use of testosterone was not associated with a significant eleva-ion in serum oestradiol. Rosati et al. studied the use of etoricoxib60 mg/day) versus placebo, in addition to anastrozole (Rosati et al.,011). This study has also only been published as an abstract. Dur-

ng the trial, there was a U.S. Food and Drug Administration (FDA)lert on the use of etoricoxib and the potential risk for cardio-ascular toxicity, resulting in a 38% discontinuation rate. Despitehis, the incidence of musculoskeletal pain was still significantlyigher in the placebo arm (RR 2.1, 95% CI 1.29–3.43, p = 0.002).he third RCT, by Liu et al. investigated calcitonin 200 IU/daylus caltrate D 600 mg/day versus caltrate D alone for a period of

hree months (Liu et al., 2014). An improvement in pain scores,

easured by VAS, was identified in both the placebo arm (scoreifference −1.00, p = 0.0013) and intervention arm (score difference−3.00, p < 0.0001). There was also notably a difference between

d deviation; MD, mean difference; CI, confidence interval.

the improvement in pain scores between the two arms (p < 0.0001)(Liu et al., 2014).

5.1. Analysis of complementary interventions

5.1.1. Study characteristicsTen studies were included that investigated the use of comple-

mentary therapies, including 6 RCTs, 2 cohort studies and 2 pre/poststudies. The interventions included Blue Citrus Herbal (Massiminoet al., 2011), omega-3-fatty-acids (O3FA) (Hershman et al., 2015;Lustberg et al., 2015), vitamin D (Shapiro et al., 2016; Rastelli et al.,2011; Prieto-Alhambra et al., 2011; Khan et al., 2010; Khan et al.,2012) vitamin E (Kiyomi et al., 2015) and glucosamine/chondroitin(Greenlee et al., 2013). The total number of patients investigatedwith complementary interventions was 403. The median samplesize was 61 (range 31–209). All RCTs used a placebo as the compara-tor arm. The primary endpoint of majority of studies included theimpact of the intervention on AIMSS. In the pilot study investigatingO3FA, the primary outcome was feasibility (Lustberg et al. 2015),but the secondary outcome included patient-reported outcomes(PRO) of AIMSS. In the pre/post study of vitamin E, the primaryoutcome was the effect of vitamin E administration on female hor-mones and cytokines in patients experiencing AIMSS (Kiyomi et al.,2015). Secondary outcomes included the effects of vitamin E onseverity of AIMSS. In the vitamin D RCTs, the dosing of vitamin D inthe intervention arms varied between studies, with interventionsincluding 4000IU/day vitamin D3 (Shapiro et al., 2016); 50,000IUvitamin D weekly (Rastelli et al., 2011); and 30,000 IU vitamin D3weekly (Khan et al., 2012).

6. Results

All three of the RCTs investigating the use of vitamin D (Shapiroet al., 2016; Rastelli et al., 2011; Khan et al., 2012), showed no ben-efit in the use of vitamin D for the management of AIMSS. The RCTinvestigating the use of Blue Citrus Herbal (Massimino et al., 2011),which included 37 patients in the study, reported improvementin VAS pain scores with the use of Blue Citrus Herbal, but it wasunclear if these improvements were statistically significant com-pared to the control arm. Some Blue Citrus herbal formulationscan include up to 15 different herbs, including Curcuma (Qi HerbalApothecary, 2017). Caution is advised for usage of Curcuma inwomen with hormone sensitive conditions as theoretically it mayexacerbate hormone sensitive breast, uterine or ovarian cancers(Natural Medicines Comprehensive Database, 2016). The specificformulation of the herb used in this trial is unclear. The two RCTsinvestigating the use of O3FA (Hershman et al., 2015; Lustberg et al.,2015) did not find any benefit of O3FA when compared to placebo.

The larger of the O3FA RCTs, comprising 249 patients (Hershmanet al., 2015), reported a substantial improvement in AIMSS at 12weeks in both the O3FA arm (BPI-SF score change −1.74, p < 0.001)and the placebo arm (BPI-SF score change −1.50, p < 0.001), but no

Oncol

sssmw

6

6

ipacpatiiOa2(atespn

7

pwtitmgrTtim

tiIes

K. Roberts et al. / Critical Reviews in

ignificant difference between groups (p = 0.38). These results wereustained at the 24 week evaluation. The only positive result in themaller O3FA RCT (Lustberg et al., 2015) was greater pain relief fromedications in the O3FA arm at both 12 weeks (p = 0.043) and 24eeks (p = 0.011).

.1. Analysis of physical therapy interventions

.1.1. Characteristics of studiesTen studies were included that investigated the use of phys-

cal therapies on the management of AIMSS, including 3 RCTs, 6re/post studies and a CCT. Three studies investigated a combinederobic and resistance exercise program, including two randomisedontrol trials (Irwin et al., 2015; Lohrisch et al., 2011), and onere/post study (Lash et al., 2011). One pre/post study investigated

home-based exercise program (DeNysschen et al., 2014), andwo studies investigated walking programs, including one RCTnvolving Nordic Walking (Fields, 2015), and one pre/post studynvestigating a self-directed walking program (Nyrop et al., 2014).ne pre/post study investigated Tai Chi (Galantino et al., 2013),nd two other pre/post studies investigated yoga (Jacobsen et al.,015; Galantino et al., 2012). A CCT investigated aquatic exerciseCantarero-Villanueva et al., 2013). Therefore, the physical ther-py interventions were extremely heterogeneous, ranging fromwo one hour tai chi sessions per week for 8 weeks (Galantinot al., 2013), to 150 min of aerobic exercise weekly plus supervisedtrength training twice weekly (Irwin et al., 2015). The mean sam-le size between studies was 31 patients (range 10–121). The totalumber of patients investigated in exercise trials was 313.

. Results

Of the 3 RCTs, there was only one study showing benefit withhysical therapy (Irwin et al., 2015). The HOPE study, by Irwin et al.,as the largest of the exercise studies in our analysis, with 121 par-

icipants (Irwin et al., 2015). The study reported a 29% improvementn worst BPI scores in the exercise group at 12 months, as comparedo a 3% increase in worst pain scores in the usual care group at 12

onths (p < 0.001) (Irwin et al., 2015). The RCT by Fields investi-ating the use of Nordic walking versus waitlist control did noteport any significant benefit in regards to AIMSS (Fields, 2015).he RCT by Lohrisch et al. (Lohrisch et al., 2011), closed early dueo poor recruitment, and did not identify any significant benefitn the use of a mixed aerobic/resistance exercise program for the

anagement of AIMSS.Two studies could be included in a meta-analysis of physical

herapy interventions, as only two studies had the same pain scor-

ng system within their studies, with available results (Fields, 2015;rwin et al., 2015) (Fig. 3). There was significant between-study het-rogeneity for the effects of physical therapy on BPI-SF worst paincore (I2 = 93%). The overall mean difference in worst pain scores

Stud y

Fixed effec t modelRando m effects mod elHeterogeneity: I -squared=93 .1%, tau-squared=4.473, p=0.000 1

FieldsIrwin

Total

81

2061

Mea n

-1.1-1.6

SD

2.002.39

ExperimentalTotal

80

2060

Mea n

-2.4 0.2

SD

2.202.77

Con trol

-2

M

Fig. 3. Effect of physical therapy on AIMSS using BPI-SF. SD, stand

ogy/Hematology 111 (2017) 66–80 73

after exercise intervention, using the random effects model was−0.29 (95% CI, −3.32–2.75).

8. Discussion

With the improving long-term prognosis for breast cancerpatients, there is an increasing focus on survivorship, and the qual-ity of life for breast cancer survivors. Despite the burden of AIMSSin the treatment of hormone receptor positive breast cancer, thereis a paucity of large, well-designed trials to provide evidence onthe management of this condition. It should be emphasized thatthere is currently no standardised definition of AIMSS. The con-dition encompasses a broad range of symptoms, and therefore astandardised definition would not only assist trial design in thefuture, but also assist oncologists to recognise and manage thiscondition in the clinical setting.

In compiling this analysis, we identified a variety of factorswhich unfortunately compromise the quality of the available evi-dence. The majority of trials considered in this analysis includedpatients who either already experienced musculoskeletal symp-toms which worsened after the initiation of an AI, or patients whodeveloped new onset musculoskeletal symptoms after initiation ofan AI. However, several trials which did not stipulate their inclu-sion requirement for AIMSS at entry. Poorly defined entry criteriamay have resulted in some trials investigating long-standing mus-culoskeletal conditions, such as osteoarthritis, rather than AIMSSspecifically. Furthermore, interventions with low perceived toxic-ity would be more likely to have uptake in patient groups with lesssevere symptoms leading to differences in patient groups betweentrials (Pavuluri et al., 2004). The retrospective studies would havelikely used physician-reported AIMSS and outcomes, whereas themost reliable process for reporting patient quality of life outcomesincludes patient reported outcomes (PRO). (Deshpande et al., 2011)

In the trials included in this analysis, a diverse range of scoringsymptoms were used to record patient symptoms. Due to the het-erogeneity of scoring systems used, it was difficult to compare thebenefit of interventions between trials. The more simplistic scor-ing systems, such as VAS, may result in either overestimation orunderestimation of the perceived benefit of an intervention. Mul-tiple studies did not disclose their complete list of scoring results,which may indicate a risk of bias in the reporting of results.

Our analysis included many studies investigating the manage-ment of AIMSS, but overall they provide poor quality evidence inthis area. The majority of studies had small sample sizes and a highrisk of inherent bias. As expected, it is extremely difficult to blindthe intervention group in certain studies, and impossible to blindtreatment arms in the physical therapy groups. The placebo effecthas been found to be significant in other studies of pharmacologi-

cal treatment of debilitating toxicities of cancer treatment (Pavuluriet al., 2004; Smith et al., 2013; Spathis et al., 2014). In addition, thechoice of placebo or control arm may have contributed to some bor-derline results. Careful consideration should be given to the choice

-1 0 1 2

ean differenceMD

-0.77-0.29

1.30-1.80

95% -CI

[-1.52; -0.01][-3.32; 2.75]

[ 0.00; 2.60 ][-2.72; -0.88]

W(fixed)

100%--

33.4%66.6%

W(rando m)

--100 %

48.9%51.1%

ard deviation; MD, mean difference; CI confidence interval.

7 Oncol

oobaswate2temsaita

vos((d

4 K. Roberts et al. / Critical Reviews in

f placebo, as contamination may be a problem. For example, inne trial investigating the use of O3FA (Hershman et al., 2015), soy-ean was used in the placebo tablet. It has been hypothesised thatn oestrogenic component in soy may have impacted on the paincores in the control arm. In a number of acupuncture trials, thereas an improvement in pain scores in both the real acupuncture

nd sham acupuncture arms. It is theorised that sham acupunc-ure may provide a therapeutic benefit by triggering the release ofndorphins or activation of pain related neural matrix (Chien et al.,015). Many of the acupuncture trials have gone to great lengthso attempt to eliminate the risk of bias from their study, but asxpected, they have not been able to successfully blind their treat-ent arms to patients. This may have resulted in a bias of results

temming from positive patient expectations. Most of the studieslso did not report trial participants’ usage of other medications,ncluding analgesia. There should be rigorous control for medica-ions taken by trial participants, to prevent confounding variablesffecting the trial outcome.

There are a number of trials currently ongoing, with inter-entions such as duloxetine (NCT01598298), the interplayf pain, sleep quality and fatigue (NCT01983995), hypno-

is (NCT02657993), testosterone (NCT01573442), acupunctureNCT01535066), vitamin D (NCT01988090), and kinesiotapingNCT02406794). Hopefully these trials will provide further evi-ence for the optimal management of AIMSS.

ogy/Hematology 111 (2017) 66–80

9. Conclusion

Suboptimal compliance with AI adjuvant therapy due to inade-quately managed AIMSS remains a major unmet need in oncologypractice. Patients who have failed to control AIMSS with over thecounter analgesics may be willing to try other interventions, includ-ing complementary therapies, for symptom relief. Many of thesewomen have been financially impacted by their cancer and itstreatment and some of the therapies discussed here may involveconsiderable financial commitment. Caution may also need to beadvised if the CAM potentially contains oestrogenic compoundswhich may explain the mechanism of action and which theoret-ically could compromise breast cancer survival. Pharmacologicaltreatment is often recommended by health professionals for AIMSS,however in conclusion, there is limited published evidence for itsuse. Exercise showed benefit in a single RCT (Irwin et al., 2015),but the other studies showed little evidence of benefit. Informationfrom the meta-analysis is limited by inclusion of only two studieswith opposing results. The evidence for acupuncture is not strongenough to recommend it for the treatment of AIMSS. Althoughthe interventions generally appear tolerable with minimal adverseeffects, the current level of evidence is low, and additional large

RCTs with more rigorous control for contamination from otherinterventions are required to confirm some of the reported promis-ing results.

Oncology/Hematology 111 (2017) 66–80 75

A

P

K. Roberts et al. / Critical Reviews in

ppendix A.

ubMed

7 Oncology/Hematology 111 (2017) 66–80

E

6 K. Roberts et al. / Critical Reviews in

MBASE

Oncology/Hematology 111 (2017) 66–80 77

C

K. Roberts et al. / Critical Reviews in

ENTRAL

7 Oncology/Hematology 111 (2017) 66–80

C

R

A

B

B

8 K. Roberts et al. / Critical Reviews in

INAHL

eferences

ydiner, A., 2013. Meta-analysis of breast cancer outcome and toxicity in adjuvanttrials of aromatase inhibitors in postmenopausal women. Breast (Edinburgh,Scotland) 22 (2), 121–129.

ao, T., Cai, L., Giles, J.T., Gould, J., Tarpinian, K., Betts, K., et al., 2013. A dual-centerrandomized controlled double blind trial assessing the effect of acupuncture inreducing musculoskeletal symptoms in breast cancer patients takingaromatase inhibitors. Breast Cancer Res. Treat. 138 (1), 167–174.

irrell, S, Tilley, W., Testosterone Undecanoate Treatment Reduces JointMorbidities Induced by Anastrozole Therapy in Postmenopausal Women withBreast Cancer: Results of a Double-Blind, Randomized Phase II Trial, 2010. 69(24 Supplement). Available from: http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/633/CN-00989633/frame.html.

Briot, K., Tubiana-Hulin, M., Bastit, L., Kloos, I., Roux, C., 2010. Effect of a switch ofaromatase inhibitors on musculoskeletal symptoms in postmenopausalwomen with hormone-receptor-positive breast cancer: the ATOLL (articulartolerance of letrozole) study. Breast Cancer Res. Treat. 120 (1), 127–134.

Burstein, H.J., 2007. Aromatase inhibitor-associated arthralgia syndrome. Breast(Edinburgh, Scotland) 16 (3), 223–234.

Cantarero-Villanueva, I., Fernández-Lao, C., Caro-Morán, E., Morillas-Ruiz, J.,Galiano-Castillo, N., Díaz-Rodríguez, L., et al., 2013. Aquatic exercise in a

chest-high pool for hormone therapy-induced arthralgia in breast cancersurvivors: a pragmatic controlled trial. Clin. Rehabil. 27 (2), 123–132.

Chien, T.J., Liu, C.Y., Chang, Y.F., Fang, C.J., Hsu, C.H., 2015. Acupuncture for treatingaromatase inhibitor-related arthralgia in breast cancer: a systematic reviewand meta-analysis. J. Altern. Complement. Med. 21 (5), 251–260.

Oncol

C

C

C

C

C

C

D

D

F

G

G

G

G

H

H

H

H

H

H

I

I

J

J

K

K

K. Roberts et al. / Critical Reviews in

oates, A.S., Keshaviah, A., Thürlimann, B., Mouridsen, H., Mauriac, L., Forbes, J.F.,et al., 2007. Five years of letrozole compared with tamoxifen as initial adjuvanttherapy for postmenopausal women with endocrine-responsive early breastcancer: update of study BIG 1–98. J. Clin. Oncol. 25 (5), 486–492.

ochrane Handbook for Systematic Reviews of Interventions: The CochraneCollaboration, 2011. Available from: www.handbook.cochrane.org.

oleman, R.E., Bolten, W.W., Lansdown, M., Dale, S., Jackisch, C., Merkel, D., et al.,2008. Aromatase inhibitor-induced arthralgia: clinical experience andtreatment recommendations. Cancer Treat. Rev. 34 (3), 275–282.

rew, K.D., Capodice, J.L., Greenlee, H., Apollo, A., Jacobson, J.S., Raptis, G., et al.,2007. Pilot study of acupuncture for the treatment of joint symptoms relatedto adjuvant aromatase inhibitor therapy in postmenopausal breast cancerpatients. J. Cancer Surviv.: Res. Pract. 1 (4), 283–291.

rew, K.D., Capodice, J.L., Greenlee, H., Brafman, L., Fuentes, D., Awad, D., et al.,2010. Randomized, blinded, sham-controlled trial of acupuncture for themanagement of aromatase inhibitor-associated joint symptoms in womenwith early-stage breast cancer. J. Clin. Oncol. 28 (7), 1154–1160.

uzick, J., Sestak, I., Baum, M., Buzdar, A., Howell, A., Dowsett, M., et al., 2010. Effectof anastrozole and tamoxifen as adjuvant treatment for early-stage breastcancer: 10-year analysis of the ATAC trial. Lancet Oncol. 11 (12), 1135–1141.

eNysschen, C.A., Burton, H., Ademuyiwa, F., Levine, E., Tetewsky, S., O’Connor, T.,2014. Exercise intervention in breast cancer patients with aromataseinhibitor-associated arthralgia: a pilot study. Eur. J. Cancer Care (Engl.) 23 (4),493–501.

eshpande, P.R., Rajan, S., Sudeepthi, B.L., Abdul Nazir, C.P., 2011. Patient-reportedoutcomes: a new era in clinical research. Perspect. Clin. Res. 2 (4), 137–144.

ields, J., 2015. A Feasibility Study of a Nordic Walking Intervention for WomenExperiencing Aromatase Inhibitor Associated Arthralgia. University ofSouthampton.

alantino, M.L., Desai, K., Greene, L., Demichele, A., Stricker, C.T., Mao, J.J., 2012.Impact of yoga on functional outcomes in breast cancer survivors witharomatase inhibitor-associated arthralgias. Integr. Cancer Ther. 11 (4),313–320.

alantino, M.L., Callens, M.L., Cardena, G.J., Piela, N.L., Mao, J.J., 2013. Tai chi forwell-being of breast cancer survivors with aromatase inhibitor-associatedarthralgias: a feasibility study. Altern. Ther. Health Med. 19 (6), 38–44.

oss, P.E., Ingle, J.N., Pritchard, K.I., Robert, N.J., Muss, H., Gralow, J., et al., 2016.Extending aromatase-inhibitor adjuvant therapy to 10 years. N. Engl. J. Med.375 (3), 209–219.

reenlee, H., Crew, K.D., Shao, T., Kranwinkel, G., Kalinsky, K., Maurer, M., et al.,2013. Phase II study of glucosamine with chondroitin on aromataseinhibitor-associated joint symptoms in women with breast cancer. Support.Care Cancer 21 (4), 1077–1087.

adji, P., Jackisch, C., Bolten, W., Blettner, M., Hindenburg, H.J., Klein, P., et al., 2014.COMPliance and Arthralgia in Clinical Therapy: the COMPACT trial, assessingthe incidence of arthralgia, and compliance within the first year of adjuvantanastrozole therapy. Ann. Oncol. 25 (2), 372–377.

enry, N.L., Banerjee, M., Wicha, M., Van Poznak, C., Smerage, J.B., Schott, A.F., et al.,2011. Pilot study of duloxetine for treatment of aromatase inhibitor-associatedmusculoskeletal symptoms. Cancer 117 (24), 5469–5475.

enry, N.L., Azzouz, F., Desta, Z., Li, L., Nguyen, A.T., Lemler, S., et al., 2012.Predictors of aromatase inhibitor discontinuation as a result oftreatment-emergent symptoms in early-stage breast cancer. J. Clin. Oncol. 30(9), 936–942.

ershman, D.L., Shao, T., Kushi, L.H., Buono, D., Tsai, W.Y., Fehrenbacher, L., et al.,2011. Early discontinuation and non-adherence to adjuvant hormonal therapyare associated with increased mortality in women with breast cancer. BreastCancer Res. Treat. 126 (2), 529–537.

ershman, D.L., Unger, J.M., Crew, K.D., Awad, D., Dakhil, S.R., Gralow, J., et al.,2015. Randomized multicenter placebo-controlled trial of omega-3 fatty acidsfor the control of aromatase inhibitor-Induced musculoskeletal pain: SWOGS0927. J. Clin. Oncol. 33 (17), 1910–1917.

owell, A., Cuzick, J., Baum, M., Buzdar, A., Dowsett, M., Forbes, J.F., et al., 2005.Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial aftercompletion of 5 years’ adjuvant treatment for breast cancer. Lancet 365 (9453),60–62.

ngle, J.N., Tu, D., Pater, J.L., Martino, S., Robert, N.J., Muss, H.B., et al., 2006.Duration of letrozole treatment and outcomes in the placebo-controlled NCICCTG MA.17 extended adjuvant therapy trial. Breast Cancer Res. Treat. 99 (3),295–300.

rwin, M.L., Cartmel, B., Gross, C.P., Ercolano, E., Li, F., Yao, X., et al., 2015.Randomized exercise trial of aromatase inhibitor-induced arthralgia in breastcancer survivors. J. Clin. Oncol. 33 (10), 1104–1111.

acobsen, P.B., Muchnick, S., Marcus, S., Amheiser, P., Reiersen, P., Gonzalez, B.,2015. Pilot study of Iyengar yoga for management of aromataseinhibitor-associated arthralgia in women with breast cancer. Psychooncology24 (11), 1578–1580.

adad, A.R., Moore, R.A., Carroll, D., Jenkinson, C., Reynolds, D.J., Gavaghan, D.J.,et al., 1996. Assessing the quality of reports of randomized clinical trials: isblinding necessary? Control. Clin. Trials 17 (1), 1–12.

adakia, K.C., Kidwell, K.M., Seewald, N.J., Snyder, C.F., Flockhart, D.A., Carpenter,

J.S., et al. (Eds.), 2016. Crossover from One Aromatase Inhibitor (AI) to Anotherin the Exemestane and Letrozole Pharmacogenetics (ELPh) Trial. ASCO AnnualMeeting Proceedings.

han, Q.J., Reddy, P.S., Kimler, B.F., Sharma, P., Baxa, S.E., O’Dea, A.P., et al., 2010.Effect of vitamin D supplementation on serum 25-hydroxy vitamin D levels,

ogy/Hematology 111 (2017) 66–80 79

joint pain, and fatigue in women starting adjuvant letrozole treatment forbreast cancer. Breast Cancer Res. Treat. 119 (1), 111–118.

Khan, Q.J., Kimler, B.F., Reddy, P.S., Sharma, P., Klemp, J.R., Fabian, C.J., 2012.Randomized trial of vitamin D3 to prevent worsening of musculoskeletalsymptoms and fatigue in women with breast cancer starting adjuvantletrozole: the VITAL trial. J. Clin. Oncol. 30 (15).

Kiyomi, A., Makita, M., Iwase, T., Tanaka, S., Onda, K., Sugiyama, K., et al., 2015.Clinical significance of female-hormones and cytokines in breast cancerpatients complicated with aromatase inhibitor-related osteoarthropathy −efficacy of vitamin E. J. Cancer 6 (4), 367–376.

Kubo, M., Onishi, H., Kuroki, S., Okido, M., Shimada, K., Yokohata, K., et al., 2012.Short-term and low-dose prednisolone administration reduces aromataseinhibitor-induced arthralgia in patients with breast cancer. Anticancer Res. 32(6), 2331–2336.

Laroche, F., Coste, J., Medkour, T., Cottu, P.H., Pierga, J.-Y., Lotz, J.-P., et al., 2014.Classification of and risk factors for estrogen deprivation pain syndromesrelated to aromatase inhibitor treatments in women with breast cancer: aprospective multicenter cohort study. J. Pain 15 (3), 293–303.

Lash, B.W., Katz, J., Gilman, P., 2011. Feasibility of a defined exercise program foraromatase inhibitor-related arthralgia (AIRA) in breast cancer survivors. J. Clin.Oncol. 29 (15).

Lintermans, A., Laenen, A., Van calster, B., Van Hoydonck, M., Pans, S., Verhaeghe, J.,et al., 2013. Prospective study to assess fluid accumulation and tenosynovialchanges in the aromatase inhibitor-induced musculoskeletal syndrome:2-Year follow-up data. Ann. Oncol. 24 (2), 350–355.

Liu, P., Yang, D.Q., Xie, F., Zhou, B., Liu, M., 2014. Effect of calcitonin onanastrozole-induced bone pain during aromatase inhibitor therapy for breastcancer. Genet. Mol. Res. 13 (3), 5285–5291.

Lohrisch, C.A., McKenzie, D., Truong, P., Jesperson, D., Gelmon, K.A., Premji, S., et al.,2011. Randomized trial of exercise versus control for musculoskeletalsymptoms from adjuvant anastrozole (A) for postmenopausal early breastcancer (PEBC). J. Clin. Oncol. 29 (15).

Lustberg, M.B., Orchard, T., Pan, X., Reinbolt, R., Logan, A., Lester, J., et al., 2015.Prevention of aromatase inhibitor (AI)-induced joint symptoms with omega-3fatty acid supplementation: a randomized placebo-controlled pilot study.Cancer Res. 75 (9).

Mao, J.J., Stricker, C., Bruner, D., Xie, S., Bowman, M.A., Farrar, J.T., et al., 2009a.Patterns and risk factors associated with aromatase inhibitor-related arthralgiaamong breast cancer survivors. Cancer 115 (16), 3631–3639.

Mao, J.J., Bruner, D.W., Stricker, C., Farrar, J.T., Xie, S.X., Bowman, M.A., et al., 2009b.Feasibility trial of electroacupuncture for aromatase inhibitor–relatedarthralgia in breast cancer survivors. Integr. Cancer Ther. 8 (2), 123–129.

Mao, J.J., Farrar, J.T., Bruner, D., Zee, J., Bowman, M., Seluzicki, C., et al., 2014.Electroacupuncture for fatigue, sleep, and psychological distress in breastcancer patients with aromatase inhibitor-related arthralgia: a randomizedtrial. Cancer 120 (23), 3744–3751.

Massimino, K., Glissmeyer, M., Wagie, T., Karamlou, K., Look, R.M., Sorenson, L.,et al., 2011. Use of Blue Citrus, a Chinese herbal remedy, to reduce side effectsof aromatase inhibitors. J. Clin. Oncol. 29 (27).

Menas, P., Merkel, D., Hui, W., Lawton, J., Harper, A., Carro, G., 2012. Incidence andmanagement of arthralgias in breast cancer patients treated with aromataseinhibitors in an outpatient oncology clinic. J. Oncol. Pharm. Pract. 18 (4),387–393.

Miller, W.R., Anderson, T.J., Evans, D.B., Krause, A., Hampton, G., Dixon, J.M., 2003.An integrated view of aromatase and its inhibition. J. Steroid Biochem. Mol.Biol. 86 (3–5), 413–421.

Muslimani, A., Spiro, T., Chaudhry, A., Taylor, H., Do, I.J., Daw, H., 2009. Aromataseinhibitor-related musculoskeletal symptoms: is preventing osteoporosis thekey to eliminating these symptoms? Clin. Breast Cancer 9 (1), 34–38.

Muss, H.B., Tu, D., Ingle, J.N., Martino, S., Robert, N.J., Pater, J.L., et al., 2008. Efficacy,toxicity, and quality of life in older women with early-stage breast cancertreated with letrozole or placebo after 5 years of tamoxifen: NCIC CTGintergroup trial MA.17. J. Clin. Oncol. 26 (12), 1956–1964.

2016. Natural Medicines Comprehensive Database. Therapeutic Research Faculty,Stockton.

Nyrop, K.A., Muss, H.B., Hackney, B., Cleveland, R., Altpeter, M., Callahan, L.F., 2014.Feasibility and promise of a 6-week program to encourage physical activityand reduce joint symptoms among elderly breast cancer survivors onaromatase inhibitor therapy. J. Geriatr. Oncol. 5 (2), 148–155.

Oh, B., Kimble, B., Costa, D.S., Davis, E., McLean, A., Orme, K., et al., 2013.Acupuncture for treatment of arthralgia secondary to aromatase inhibitortherapy in women with early breast cancer: pilot study. Acupunct. Med. 31 (3),264–271.

Olivo, S.A., Macedo, L.G., Gadotti, I.C., Fuentes, J., Stanton, T., Magee, D.J., 2008.Scales to assess the quality of randomized controlled trials: a systematicreview. Phys. Ther. 88 (2), 156–175.

Partridge, A.H., LaFountain, A., Mayer, E., Taylor, B.S., Winer, E., Asnis-Alibozek, A.,2008. Adherence to initial adjuvant anastrozole therapy among women withearly-stage Breast cancer. J. Clin. Oncol. 26 (4), 556–562.

Pavuluri, M.N., Graczyk, P.A., Henry, D.B., Carbray, J.A., Heidenreich, J., Miklowitz,D.J., 2004. Child- and family-focused cognitive-behavioral therapy for pediatric

bipolar disorder: development and preliminary results. J. Am. Acad. ChildAdolesc. Psychiatry 43, 528–537.

Presant, C.A., Bosserman, L., Young, T., Vakil, M., Horns, R., Upadhyaya, G., et al.,2007. Aromatase inhibitor-associated arthralgia and/or bone pain: frequency

8 Oncol

P

Q

R

R

R

S

Gosomanidis, D., et al., 2010. New views on treatment of aromatase inhibitorsinduced arthralgia. Eur. J. Cancer 8 (Suppl. 3), 164–165.

Zhang, Q., Tang, D., Zhao, H., 2010. Immunological therapies can relieve aromatase

0 K. Roberts et al. / Critical Reviews in

and characterization in non-clinical trial patients. Clin. Breast Cancer 7 (10),775–778.

rieto-Alhambra, D., Javaid, M.K., Servitja, S., Arden, N.K., Martinez-Garcia, M.,Diez-Perez, A., et al., 2011. Vitamin D threshold to prevent aromataseinhibitor-induced arthralgia: a prospective cohort study. Breast Cancer Res.Treat. 125 (3), 869–878.

i Herbal Apothecary: The Natural Choice − Blue Citrus Tallahassee, Fl, U.S.A.: QiHerbal Apothecary; [Available from: http://qiherbs.com/blue-citrus.html.

astelli, A.L., Taylor, M.E., Gao, F., Armamento-Villareal, R., Jamalabadi-Majidi, S.,Napoli, N., et al., 2011. Vitamin D and aromatase inhibitor-inducedmusculoskeletal symptoms (AIMSS): A phase II, double-blind,placebo-controlled, randomized trial. Breast Cancer Res. Treat. 129 (1),107–116.

egan, M.M., Neven, P., Giobbie-Hurder, A., Goldhirsch, A., Ejlertsen, B., Mauriac, L.,et al., 2011. Assessment of letrozole and tamoxifen alone and in sequence forpostmenopausal women with steroid hormone receptor-positive breastcancer: the BIG 1–98 randomised clinical trial at 8·1 years median follow-up.Lancet Oncol. 12 (12), 1101–1108.

osati, M.S., Di Sri, M., Baciarello, G., LO Russo, V., Grassi, P., Marchetti, L., et al.,2011. Etoricoxib and anastrozole in adjuvant early breast cancer: ETAN trial

(phase III). J. Clin. Oncol. 29 (15).

anta-Maria, C.A., Bardia, A., Blackford, A., Connolly, R.M., Fetting, J.H., Jeter, S.,et al., 2014. A phase II study evaluating efficacy of zoledronic acid inprevention of aromatase inhibitor (AI)-associated musculoskeletal symptoms:the ZAP trial. J. Clin. Oncol. 32 (15).

ogy/Hematology 111 (2017) 66–80

Schwarzer, G., 2016. Meta|: General Package for Meta-Analysis. R Package Version4. 4-1. https://CRAN.R-project.org/package=meta.

Sestak, I., Sapunar, F., Cuzick, J., 2009. Aromatase inhibitor-induced carpal tunnelsyndrome: results from the ATAC trial. J. Clin. Oncol. 27 (30), 4961–4965.

Shapiro, A.C., Adlis, S.A., Robien, K., Kirstein, M.N., Liang, S., Richter, S.A., et al.,2016. Randomized, blinded trial of vitamin D3 for treating aromataseinhibitor-associated musculoskeletal symptoms (AIMSS). Breast Cancer Res.Treat. 155 (3), 501–512.

Smith, E.M.L., Pang, H., Cirrincione, C., Fleishman, S., Paskett, E.D., Ahles, T., et al.,2013. Effect of duloxetine on pain, function, and quality of life among patientswith chemotherapy-induced painful peripheral neuropathy: a randomizedclinical trial. JAMA 309 (13), 1359–1367.

Spathis, A., Fife, K., Blackhall, F., Dutton, S., Bahadori, R., Wharton, R., et al., 2014.Modafinil for the treatment of fatigue in lung cancer: results of aplacebo-controlled, double-blind, randomized trial. J. Clin. Oncol. 32 (18),1882–1888.

Xepapadakis, G., Ntasiou, P., Koronarchis, D., Koufoudakis, D., Panousis, D.,

inhibitor-related joint symptoms in breast cancer survivors. Am. J. Clin. Oncol.33 (6), 557–560.