LAUSR

We read with interest the recent article by Kleckner et al. [1] examining the effects of a 6-week exercise program (EXCAP©®) among 355 cancer survivors (breast, lymphoma, colon, lung, and other) on chemotherapy-induced peripheral neuropathy (CIPN). We agree that CIPN is an important patient outcome in the exercise and cancer field. In their trial, the authors concluded Bexercise appears to reduce CIPN symptoms in patients receiving taxane-, platinum-, or vinca alkaloid-based chemotherapy.^ (p. 1019). In our opinion, while the study shows potentially positive trends, the conclusions drawn by the authors may be somewhat premature. The authors stated, BExercise reduced CIPN symptoms of hot/coldness in hands/feet (-0.46 units, p=0.045) and numbness and tingling (-0.42 units, p=0.061) compared to the control. Exercise reduced CIPN symptoms more for patients who were older (p=0.086), male (p=0.028), or had breast cancer (p=0.076).^ (p. 1019). Table 2 and Figure 2 in the Kleckner et al. article show that exercise did not reduce these symptoms, and instead these symptoms increased. It appears the intervention may have slightly attenuated an increase in symptoms in the exercise group, as increases in this group were smaller in magnitude compared to the control group (although the clinical relevance of this attenuation is tenuous). While an attenuation of symptoms is a promising finding, we propose that the authors consider revising their conclusions to more accurately reflect and describe the data and the results (i.e., symptoms increased, but there was a small attenuation of symptom exacerbation in the exercise group). In their analyses comparing the two groups on CIPN outcomes, the authors did not adjust for multiple comparisons. This should be considered when interpreting the results of the study, given the alpha 0.05 threshold is for a single comparison and the likelihood of chance findings increases as more comparisons aremade. If a Bonferroni correctionwere applied to the two primary outcomes, an alpha of 0.025 would be the threshold for determining significance (0.05/2 = 0.025), and findings presented would not meet the threshold for significance. For the subgroups analyses, adjusted findings would not approach a trend signal. If a case is to be made for refraining to adjust for multiple comparisons, other strategies indicating the magnitude of effects should be presented to assist the reader in reaching a reasonable conclusion, such as an effect size [2]. The authors indicated the clinical significance of these effects is likely; however, no evidence to support this contention is presented. To measure CIPN, the authors used two single-item indicators of (a) numbness and tingling, and (b) hot/coldness in hands/feet. Two citations for these items were provided [3, 4] and the authors indicated, Bvalidity and reliability have been demonstrated for similar scales of numbness and tingling for cancer patients.^ (p. 1021). Both studies cited examined breast cancer survivors and did not evaluate an item referring to Bhot/ coldness in hands/feet.^ Caution must be taken when administering items validated in other tumor groups for reasons including differing treatment protocols and treatment plans. Comprehensive self-report scales that cover the spectrum of CIPN, with demonstrated reliability and validity, are available for use (e.g., FACT-Taxane [5]). Despite single-item indicators being suspect (e.g., unknown internal consistency, random measurement errors), we recognize the utility of using single-item indicators when evidence suggests the item possesses some degree of predictive validity with respect to the broader construct (in this case, CIPN). Nonetheless, * J.K. Vallance [email protected]