Upper-body morbidity following breast cancer treatment is common, may persist longer-term and adversely influences quality of life

This work forms a planned component of a broader study, The Pulling Through Study, which was designed to document the prevalence and severity of problems, including UBF, between six- and 18-months following breast cancer surgery among Australian women [21]. Women aged 74 years or younger, with a first diagnosis of invasive, unilateral breast cancer, and residing within a 100 kilometre radius of Brisbane, Queensland, were eligible for participation. Eligibility criteria minimised the impact of age-related co-morbidities on study findings, allowed for the untreated side to serve as a 'control' for certain outcomes, and enabled attendance at regular clinical assessments. Younger women (< 50 years) were over-sampled to ensure adequate numbers were available for specific age group analyses. Population-based sampling was undertaken through the Queensland Cancer Registry. It takes three to four months from the point of cancer diagnosis for patient information to arrive at the registry. Consequently, recruitment procedures commenced around four months PS. Following ethical approval (Queensland University of Technology, Ref No 2179H), doctor consent was obtained for 417 women (82% of random sample). Participant consent was obtained from 71% (n = 294); thereafter, seven women decided not to participate or could not be recontacted. Hence, 287 women participated in baseline measures; the majority participated in all components of data collection (75%), while the remainder participated on a 'questionnaire-only' basis.

Clinical assessment of UBF and completion of patient-administered questionnaires occurred every three months, over a 12-month period, with baseline assessment at six-months PS. Disease characteristics were collected from pathology reports at the Queensland Cancer Registry. Personal characteristics (such as side of dominance and income) were self-reported, and lymphoedema status was evaluated objectively using bioimpedance spectroscopy [21].

Clinical assessments of UBF were conducted for strength and endurance, hand grip strength, and flexibility, in that order. Upper-body strength and endurance (UBSE) were measured using an incremental exercise protocol, with each stage lasting one minute in duration and increments made by increasing speed of movement and weight held (0.5 kilogram increments, with the first one-minute stage commencing with no weight held). The movement combined a traditional 'upright row' and 'shoulder press', but the specific range of movement was individualised for each participant and each arm. To advance levels, the participant must have maintained correct form, range of movement and speed for the entire one-minute stage. Weight (kilograms, kg) held during the last successfully completed stage, assessed separately for each arm, was recorded. More details including comparison of this technique with assessment of strength and endurance using an isokinetic dynamometer are reported elsewhere [15]. A standard calibrated hand dynamometer (TTM Original Dynamometer 100 kg, Tokyo) was used to measure isometric hand grip strength (HGS). Three maximal contractions on each side, alternating between sides to allow for rest between each contraction, were performed and the maximum score achieved for each hand was recorded. Goniometry of flexion of the shoulder was used as a measure of shoulder flexibility. The participant was asked to hold her arm by her side, palm of hand positioned medially. The testing arm was then moved upward in the sagittal plane (in the flexion direction), as far as possible, with the palm of the hand always facing medially and elbow joint extended at 180 degrees. The test was undertaken twice for each side, alternating between sides. A third measure was taken when the first two measures differed by more than five degrees. The maximum range achieved for each arm was recorded. For all objective measures, results from the treated side are presented here.

The Disability of the Arm, Shoulder and Hand (DASH) questionnaire was used to assess self-reported UBF. The DASH [22] comprises 30 items and collects information about the level of difficulty experienced when performing specific tasks, the extent to which any upper-body problem interferes with normal activities, and the severity of specific upper-body symptoms. Final scores range from 0 to 100, where 0 reflects no disability (good function) and 100 reflects extensive disability (poor function).

The Functional Assessment of Cancer Therapy-Breast (FACT-B+4) questionnaire was included in the self-administered survey, to provide a measure of QoL. Final scores range from 0 to 160 whereby 0 reflects low QoL and 160 reflects high QoL. This QoL tool has been widely used and extensively described, including its validity and reliability, by others [23].

Distributions of UBF and QoL measures were approximately Normally distributed, hence means and standard deviations (SD) were used to summarise data at each phase. Absolute change in UBF between six- and 18-months PS was calculated and used to categorise participants as experiencing declines, no change, or improvements in UBF over time. Clinically, we argue that improvement following surgery is expected, with any decline in function over time potentially important, although we acknowledge that measurement error may account for some misclassification of decliners. In a sensitivity analysis of this strict position, we also categorised decline according to changes of more than 10% and reanalysed. Binary logistic regression was used to explore characteristics associated with declines in function, while regression models utilising a general linear modelling framework were used to investigate predictive relationships between function and QoL. Final regression models presented include adjustment for all potential confounders identified by statistical and/or clinical significance (further details regarding modelling provided in Hayes et al [21]). Differences of seven units in the FACTB+4 total score [23] and odds ratios (OR) of > 2 or < 0.5 were considered a priori as potentially clinically important, while a two-tailed p < 0.05 was considered statistically significant. Means (95% confidence interval, CI) and odds ratios (95% CI) are presented for continuous and dichotomous outcomes, respectively. Underlying assumptions for these analytical techniques, including the absence of multi-collinearity, were tested and met.

Overall, UBF improved throughout the 12-month testing period, with the majority of improvement occurring between six- and 12-months PS, irrespective of whether UBF was measured objectively or via self-report (Table 1). Depending on the UBF measure, up to 42% of women experienced UBF declines between six- and 18-months PS (Table 2). These results were insensitive to classification of declining UBF as either any decline or decline > 10%; classifications were unchanged for UBSE (data not shown) and similar for DASH (28% instead of 34% were categorised as decliners).

Adjusted relationships between personal and treatment characteristics and odds of decline in UBF between six- and 18-months PS are presented in Table 3. Older age, lower socioeconomic status as defined by income, being treated on the dominant side, having a mastectomy and more extensive lymph node surgery were associated with 2- to 3.5-fold increased odds of decline in objective UBF, with the relationships supported statistically (p < 0.05) except for having a mastectomy and extent of lymph node removal. Lower socioeconomic status, being treated on the dominant side and more extensive lymph node removal were also associated with increased odds of self-reported UBF declines (OR: 2.2-5.4; p < 0.05). In addition, older age was statistically but not clinically (OR: 1.9, p = 0.05) associated with increased odds of self-reported UBF declines, while having lymphoedema was clinically but not statistically associated with declines (OR: 2.4, p = 0.21).

Cross-sectional analyses using data from six- and 18-months PS demonstrate that there is a strong, inverse association between self-reported UBF (lower scores indicate better function) and QoL (r = -0.7 at each time, p < 0.01), but weak linear association between objective UBF and QoL (r = 0.18, p < 0.05 at six-months PS; r = 0.1, p = 0.12 at 18-months PS). Table 4 presents results of linear regression analyses exploring the relationships between change in UBF (from six- to 18-months PS) and QoL at 18-months PS, adjusting for baseline UBF as well as potential confounders identified elsewhere [24]. Declines in UBF between six- and 18-months PS were associated with lower QoL at 18-months PS, but the association was only clinically and statistically significant for self-reported UBF (greater than 7 unit difference between decliners and improvers, p < 0.01).