LAUSR

Dear Mr. Ashbury, We are writing to seek clarification and request correction of errors in the article “The effect of exercise on aromatase inhibitor-induced musculoskeletal symptoms in breast cancer survivors: a systematic review and meta-analysis” by Lu et al. [1] published on line in your journal on 18 December 2019. The analysis is incorrect as Lu et al. [1] have included patients from the HOPE study [2] three times in their analysis. In separate outcome/substudy reports of the original HOPE study, Irwin et al. [2] have also been included under reports from Baglia [3] and Thomas [4] as if they were independent studies, when this is not the case. The “Characteristics of the included studies” table displays this with three of the nine included studies having identical numbers of patients (N = 121 each). This should have been an alert flag of possible error, even if the studies were not reviewed correctly. The statistical and resultant conclusions will therefore not be correct. The review also cites the study by Cantarero-Villaneuva [5] as a randomized controlled trial which is incorrect. Cantarero-Villaneuva et al. [5] state that they were unable to randomize patients for “ethical reasons.” We also have concerns about the eligibility of the included participants. More than 50% of patients in the study by Cantarero-Villaneuva [5] were treated with tamoxifen. Similarly, 45% of patients in the study byNikander [6] received tamoxifen. No commentary is provided on how the authors dealt with these ineligible populations within these included studies. Furthermore, the authors state that they were following the principles of the PRISMA statement to write their review, yet they have not provided a complete and reproducible search strategy for at least one of the databases. We have recently published a Cochrane review “Exercise therapies for preventing or treating aromatase inhibitor-induced musculoskeletal symptoms in early breast cancer (Review)” [7] in follow-up to our previously published Cochrane protocol on this topic [8]. Seven studies with 400 randomized participants were included. In view of the widely acknowledged benefits of exercise in cancer, it was surprising that we found no clear evidence of benefit of exercise in AIMSS in early breast cancer [7]. However, the quality of the evidence was of low and very low certainty in the various outcomes. The studies by CantareroVillaneuva [5] and Nikander [6] were excluded from our review for reasons outlined above. We request clarification from the study authors of Lu et al. [1] regarding the significant analytic errors and lack of appropriate screening of the content of the included studies. Yours sincerely, Dr. Kate Roberts Kirsty Rickett Dr. Sophie Feng Dimitrios Vagenas Assoc Prof Natasha Woodward