This study aimed to report the relationship between changes in subjective shoulder disability and various shoulder functions. Although the variables regarding exercise and rehabilitation were excluded from the covariates, initiation of exercise should be considered as early as possible, and practice of stretching exercises should be continued for subjective improvement. When the effect of chemotherapy history and reconstruction type were statistically controlled, the abduction improvement and neuropathic pain decrease were significantly correlated with subjective improvements in upper limb functionality. Despite the lack of an interaction effect between them, a trend for interaction was observed (Figure 2). The present study reported no adverse effects of chemotherapy following IBR; however, subjective upper limb function might not improve.
The analysis using this increase was different from the cross-sectional analysis (Supplementary Tables S3 and S5). The main result mostly agrees with a similar analysis [21], while it partially agrees with cross-sectional studies [13, 20]. Pectoral tightness was confirmed by observation. In this study, the mean shoulder ROM was a <150-degree elevation with a >80-degree rotation after surgery. Given that abduction requires adequate scapular adduction and humeral external rotation, the pectoralis major tightness appears to be key in the limitation of arm abduction. The improvement in muscle tightness is directly related to amelioration in upper limb disability and indirectly related to QoL improvement via PF, RF, and BS (Supplementary Table S2). Given that 10 PMI was regarded as normal in the resting position [35], the pectoralis minor length was not significantly shortened after IBR. As reported, the pectoralis minor was not severely shortened after surgery, and returned normal at 4 months postoperatively. Thus, pectoralis major tightness should be measured, and its acute rehabilitation should be prioritized after IBR, rather than that of the pectoralis minor.
The neuropathic pain score change was also the most significant variable for the Q-DASH score change, which agrees with a previous report that neuropathic pain was the sole significant variable explaining the DASH score. In this study, this score change correlation coefficient was slightly larger (r=0.49) than the correlation coefficient of the PA change from the EORTC C30 (r=0.47) to the improvement in the Q-DASH score, and the two pain score changes were moderately correlated with each other (r=0.46). The major finding of this study was that the improvement in the PDQ score was independent of the abduction and strength changes (Supplementary Table S1). A recent systematic review article [36] recommended the use of transcutaneous electrical nerve stimulation (TENS) and physiotherapy in conjunction with pharmaceutical or cognitive intervention, but there was no consensus on the appropriate physiotherapy. Given that manual therapy, including massage, has a positive effect on postoperative pain in BrCS [37], adequate physiotherapy might provide TENS and manual therapy. While until now rehabilitation programs have commonly aimed at ROM improvement, future rehabilitation programs will preferentially aim to manage the painful nature of the condition rather than the ROM. However, further research must be conducted to shed light on this issue.
The improvement in dynamic shoulder strength was not related to the improvement in other shoulder functions and subjective reports. Although the cross-sectional relationship between strength and the Q-DASH score (Supplementary Table S5) was similar to that reported by Harrington and colleagues [20], the increase in strength was not similar. Myung and colleagues [7] described that muscle continuity provided better exercise adherence for the full recovery of shoulder strength. Given the lack of a relationship between the ROM and strength increases, the shoulder strength increase seems to be a natural recovery rather than due to the muscle length change. However, the strength index can be used as a predictor. According to the prediction analysis (Supplementary Table S8), the early peak torque of rotation significantly predicted late upper-limb disability. Therefore, shoulder strengthening is not necessary for acute postoperative rehabilitation in this population, but the initial measurement could distinguish rehabilitation candidates.
Similarly, shoulder proprioception improvement was not related to upper limb disability. Given the possible relationship between movement accuracy and social function, this function should not be ignored. Considering that social function could be influenced by upper-limb disability, movement quality should be measured and targeted, especially in patients with BrCS who underwent breast reconstruction. A previous study reported that this surgery may alter the shoulder biomechanics [8]; hence, the movement quality was poorer than that of the BrCS who underwent mastectomy and controls. In addition, the present study reported a possible explanation value of shoulder proprioception improvement for the improvement in the upper limb disability score. Despite insignificance (p<0.07), the improvement in shoulder proprioception could explain 10% of the Q-DASH score change in the BrCS without CTx (Table 4). In a previous study, movement quality, pain intensity, and DASH score improved after rehabilitation, including pectoral muscle stretching [9]. In this study, the POD of the participants was approximately 18 months, and approximately 40% of the participants received chemotherapy. Considering that systemic therapy ends within a year after surgery, the negative effects of chemotherapy might decrease. Therefore, the results of the previous study may align with the results of the present one. When the BrCS were not administered chemotherapy after IBR, movement accuracy could be improved in the rehabilitation program to ameliorate upper limb disability.
As the accuracy of LRJT improved throughout the cohort, a positive relationship with upper limb disability change was expected. However, there was no relationship with upper limb disability. We only found a trend for a positive relationship between the improvement of body schema and the improvement of social function (Supplementary Table S2) and confirmed the predictive value of LRJT accuracy (Supplementary Table S8), as we reported previously [15]. Brain-targeted intervention, such as Graded Motor Imagery (GMI), to improve LRJT results has been recently studied in various pain conditions. In addition to studies that included participants with chronic regional pain syndrome, GMI was only applied to participants with chronic shoulder pain [38], frozen shoulder [39, 40], and knee osteoarthritis [41]. Only two studies implemented the LRJT [40, 41], but the effect of GMI was controversial between two results. Given this controversy, the present result could be explained by the fact that we provided a rehabilitation program without GMI. As the central mechanism for pain control was not manipulated by a specific regimen such as GMI, the degree of pain and disability was sustained in this study. Thus, improvement in WBS did not change upper limb disability or pain in this population. Along with neuropathic pain rehabilitation, this is a challenge to be addressed in the future.
This study reported both subjective and objective outcomes of BrCS who underwent IBR. Given the increasing importance of early rehabilitation after IBR, this study might help clinicians understand the physical characteristics and impact of subjective reports. In addition, we captured the real course of breast cancer treatment, in which systemic therapy or local therapy began after surgery, while early mobilization started as soon as possible.
However, this study should be interpreted with caution. First, the small sample size reduced the power of the analysis. Second, a few variables were not adequately changed. While the LRJT accuracy was statistically improved, the difference was lower than the previous minimal clinically important difference of 10%. Third, the measurement method may have decreased study quality. The length of the pectoralis minor was palpated rather than imaged by X-ray; the absolute angle error was modified from the standard position, and the test-retest reliability was approximately 0.7 [15]. Fourth, the data on non-protocol exercises such as yoga, which may have affected upper limb function, were not recorded. Fifth, the cause-effect relationship was not clear owing to differences in cancer therapy and rehabilitation therapy. Finally, there was no normative Q-DASH score for the BrCS, and we could not identify the normal value of each function regarding the Q-DASH score.