Discussion

The prevalence of patients with arm lymphedema after breast cancer treatment was 6·2%. While several studies have reported this prevalence ranges between 3% and 70%, comparisons across studies are difficult because of variability in arm lymphedema definition and measurement techniques, follow-up times, and patient characteristics. ^{Reference DiSipio, Rye, Newman and Hayes13–Reference Ribeiro Pereira, Koifman and Bergmann19} In a recent meta-analysis, the overall estimated incidence of the condition after breast cancer treatment was 16·6%. ^{Reference DiSipio, Rye, Newman and Hayes13} The reason for such a low prevalence rate in our study may have been due to the fact we used only one method of arm lymphedema assessment. From other studies, incidences were high when assessed by more than one diagnostic method ^{Reference DiSipio, Rye, Newman and Hayes13} and also from a lack of diagnostic criteria for breast cancer-related arm lymphedema. ^{Reference DiSipio, Rye, Newman and Hayes13} Another reason for a possible underestimation may have been due to the median follow-up time; in this study, this was 4·4 years, thus the longer a woman’s exposure to risk factors, the greater the chance of developing arm lymphedema. With a 20-year follow-up, Petrek et al. ^{Reference Coen, Taghian, Kachnic, Assaad and Powell25} reported that the ongoing risk of developing the condition was approximately 1% per year for at least 20 years.

Current data on whether arm lymphedema incidence varies by age are inconsistent. A higher incidence of arm lymphedema in older patients was observed in some studies, ^{Reference Bergmann, Mattos, Koiffman and Ribeiro16,Reference Ugur, Arıcı and Yaprak23,Reference Ahmed, Schmitz, Prizment and Folsom24,Reference Clough-Gorr, Ganz and Silliman27} possibly due to a progressive loss of lymphvenous anastomosis due to ageing processes. ^{Reference Deo, Ray and Rath17–Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Ribeiro Pereira, Koifman and Bergmann19–Reference Norman, Localio and Kallan28} This finding was not observed in our study; patients >60 years old showed no increased probability of developing arm lymphedema after adjusting for others characteristics.

BMI is a modifiable risk factor for arm lymphedema; previous studies have shown that patients with a higher BMI were at a higher risk of arm lymphedema. ^{Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Ahmed, Schmitz, Prizment and Folsom24,Reference Petrek, Senie and Peters26,Reference Norman, Localio and Kallan28,Reference Ridner, Dietrich, Stewart and Anmer30} The mechanism whereby excess weight increases the risk of the condition remains unclear; however, extra demands on both the vascular and lymphatic system to transport fluid could impair lymphatic transport capacity and impair lymphatic functions thereby promoting adipose deposition. ^{Reference Ridner, Dietrich, Stewart and Anmer30,Reference Mehrara and Greene31} In contrast, our study failed to show an association between higher BMI and increased risk of arm lymphedema development; therefore, more studies are required to explore such associations.

The majority of studies show that arm lymphedema risk increases with more extensive treatments, in particular breast surgery, ^{Reference DiSipio, Rye, Newman and Hayes13,Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Coen, Taghian, Kachnic, Assaad and Powell25–Reference Norman, Localio and Kallan28} lymph node removal ^{Reference Ahmed, Schmitz, Prizment and Folsom24,Reference Paskett, Naughton, McCoy, Case and Abbott32,Reference Kwan, Darbinian and Schmitz33} and adjuvant treatment. ^{Reference Ragaz, Jackson and Le7,Reference DiSipio, Rye, Newman and Hayes13,Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Jung, Shin and Kim21,Reference Ahmed, Schmitz, Prizment and Folsom24,Reference Paskett, Naughton, McCoy, Case and Abbott32,Reference Kwan, Darbinian and Schmitz33} Several studies have demonstrated arm lymphedema rates of 24% to 49% after mastectomy and 4% to 28% after lumpectomy, with a history of axillary lymph node surgery. ^{Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Coen, Taghian, Kachnic, Assaad and Powell25–Reference Norman, Localio and Kallan28} The Axillary Lymphatic Mapping Against Nodal Axillary Clearance trial demonstrated that lymphedema was observed in approximately 5% of patients who had SLNB versus 13% of those who had ALND. ^{Reference Mansel, Fallowfield and Kissin34} Radiation therapy is an independent risk factor for the development of arm lymphedema with reported rates of 2% to 5%, even in the absence of lymph node surgery. ^{Reference Kiel and Rademacker35,Reference Johansen, Overgaard, Blichert-Toft and Overgaard36} Radiation to the axilla is associated with 2 to 4·5 times greater risk of lymphedema and 8 to 10 times greater risk when a patient receives ALND and radiation treatment. ^{Reference Kissin, Querci Della Rovere, Easton and Westbury37,Reference Kwan, Jackson and Weir38} For systemic treatment, there are conflicting studies about chemotherapy as possible risk factors for arm lymphedema. Some studies indicate that adjuvant chemotherapy is a potential risk factor for arm lymphedema. ^{Reference Ragaz, Jackson and Le7,Reference DiSipio, Rye, Newman and Hayes13,Reference Ribeiro Pereira, Koifman and Bergmann19,Reference Jung, Shin and Kim21,Reference Ahmed, Schmitz, Prizment and Folsom24,Reference Paskett, Naughton, McCoy, Case and Abbott32,Reference Kwan, Darbinian and Schmitz33} DiSipio et al. ^{Reference DiSipio, Rye, Newman and Hayes13} did a meta-analysis of 72 studies published from 2000 to 2012 to assess the incidence rates of BCRL, and a meta-analysis of 29 studies published from 2000 to 2012, to assess the risk factors. This study demonstrated that about 75% of the studies reviewed indicated chemotherapy as a risk factor for developing arm lymphedema. ^{Reference DiSipio, Rye, Newman and Hayes13} Our study also demonstrated that having >20 lymph nodes removed during axillary dissection increased the HR 3.3-fold, and having regional lymph node irradiation increased the HR 1.81-fold. We failed to confirm an association between arm lymphedema development and type of breast surgery, type of lymph node dissection and adjuvant systemic treatment. The reason for this may be due to improvements in treatment techniques to minimise lymphatic disruption and the improved selection of systemic treatments in recent years. ^{Reference Shaitelman, Cromwell and Rasmussen39}

From our multivariate analyses, staging at diagnosis did not increase the risk of developing arm lymphedema. This agreed with the systematic review; staging was classified as a weak or inconclusive independent risk factor for developing arm lymphedema. ^{Reference DiSipio, Rye, Newman and Hayes13}

Arm and shoulder exercise after breast cancer treatment is an important independent risk factor for arm lymphedema. We observed that arm and shoulder exercises decreased the risk of developing arm lymphedema by 0·08 and 0·15 times, based on univariate and multivariate analyses, respectively. Consistent with others study, arm and shoulder exercises reduced breast cancer treatment complications and side effects, including arm lymphedema, functional limitations and upper limb disability. ^{Reference Courneya, Segal and Mackey40–Reference Milne, Wallman, Gordon and Courneya43}