ABSTRACT: Lymphedema is a feared complication of cancer treatment and one that negatively impacts survivorship. The incidence of breast cancer–related lymphedema ranges from 6% to 70%, but lymphedema may be a common and under-reported morbidity. No standard guidelines for its diagnosis and assessment exist. Although the true etiology of lymphedema remains unknown, radiation, chemotherapy, type of breast surgery, and extent of axillary surgery are commonly cited risk factors. However, the relationship between the number of nodes removed and the risk of lymphedema is not clearly correlated. Clinical trials are focusing on ways to reduce the need for axillary dissection even in the setting of a positive sentinel node, to help minimize axillary morbidity. Risk-reduction practices, including avoidance of skin puncture and blood pressures in the ipsilateral upper extremity, and precautionary behaviors such as wearing compression garments during air travel continue to be advocated by the medical and survivor communities, despite a lack of rigorous evidence supporting their benefit. Emerging data support exercise in at-risk and affected women with lymphedema when started gradually and increased cautiously.
The most recent survivorship data from the Surveillance, Epidemiology, and End Results (SEER) database demonstrate a significant increase in the number of US cancer survivors over the last 30 years, from roughly 3 million in 1973 to nearly 12 million in 2008. As a result, issues of survivorship have stimulated new focus for clinical trials, not only to determine the most effective therapeutic regimen (surgery, drug, or radiation) but also to identify the one with the least influence on future quality of life. Lymphedema has long been a feared complication of surgical cancer treatment, and notably one that negatively impacts survivorship. Fear of lymphedema stems from patient concerns regarding the chronic, progressive nature of the condition and the clinician’s relative inability to predict or prevent its development. Furthermore, decades of physician and allied health teachings based on opinion and theory have perpetuated the myths shrouding lymphedema risk, prevention, and treatment.
A vast body of literature documents the occurrence of breast cancer–related lymphedema, with more than 1400 articles indexed in PubMed-MEDLINE databases alone. Importantly, lymphedema also exists after surgery for non–breast-cancer-related malignancies, but data documenting this occurrence are rare in comparison. Recently, Cormier et al found only 47 studies between 1972 and 2008 with more than 10 patients that prospectively evaluated lymphedema as a primary or secondary outcome after treatment for melanoma, bladder, sarcoma, penile, prostate, vulvar, cervical, endometrial, or head and neck cancers. The authors’ analysis of these studies demonstrated the overall incidence of lymphedema to be 16.3% after melanoma, 10.1% after genitourinary cancers, and 19.6% after gynecologic malignancies, and notes that lymphedema rates are higher when the lower rather than upper extremity is affected. Given the abundance of breast cancer data, this review will focus on breast cancer–related lymphedema. However, the principles and controversies discussed are relevant regardless of the type of malignancy to which the lymphedema is attributed.
Incidence and Diagnosis
The incidence of breast cancer–related lymphedema has been difficult to quantify due to delayed onset of symptoms and the lack of standardized diagnostic criteria. A recent meta-analysis reports the incidence of breast cancer–related lymphedema to range from 0 to 3% after lumpectomy alone to as high as 65% to 70% after modified radical mastectomy (removal of breast and axillary lymph nodes) with regional nodal radiation. Overall, 80% to 90% of women who will develop lymphedema do so within 3 years of treatment,[4,5] but the risk persists years later as the remaining 10% to 20% will develop lymphedema at a rate of 1% per year. Petrek et al followed 263 patients and found that nearly 50% developed lymphedema by 20 years. These data suggest that lymphedema is probably more common than generally reported, and clearly the length of follow-up in a given study influences the reported incidence. A uniform definition of lymphedema does not exist. Although volume displacement methods are considered the gold standard for diagnosis, these methods are cumbersome and unable to identify subclinical lymphedema. Thus, displacement methods are rarely used in clinical settings. Circumferential arm measurements performed with a nonelastic tape measure are commonly used to determine upper extremity size differences. It is essential to obtain baseline preoperative measurements of the ipsilateral and contralateral extremities, as differences of up to 2 cm may naturally exist between the dominant and nondominant arms at baseline. Published literature demonstrates significant heterogeneity, as studies lack agreement on the location and minimal number of measurements performed. Furthermore, they use multiple definitions of what constitutes lymphedema, including measurement changes greater than 2 cm, volume increase of more than 200 mL, or percentage increases in volume (> 10% or > 20%) compared with baseline when controlled for changes in the contralateral arm.[7,8] Aside from objective measurements, some studies define lymphedema by patient subjective symptoms or patient perceptions of swelling.
Controversy remains as to whether a measurement change or a patient’s perception of swelling should be considered diagnostic. While a measurement change is objective, measurements may be affected by inter-rater and even intra-rater variability affecting reproducibility. Perhaps more importantly, a defined measurement change may be asymptomatic in an obese patient, while a thin woman may be bothered by more subtle measurement changes not meeting measured criteria for lymphedema. Alternatively, a patient’s perception of swelling may more accurately diagnose symptomatic lymphedema, but this perception may be influenced by sensory changes from the surgery and additional adjuvant therapies. Further, women with surgery in their dominant arm may perceive functional issues to a greater degree than if it were the nondominant arm that was affected. In fact, one study found only 41% of women with perceived swelling had circumferential arm measurement changes greater than 2 cm, the objective threshold used in the study to define lymphedema. Several contemporary studies have evaluated the differences between measured and perceived lymphedema and have validated patient perceptions against physical therapist–directed measurements.[4,10] These studies conclude that each evaluation method is valid but not interchangeable, and they do not endorse one method over another.
The clinical introduction of a single-frequency bioelectrical impedance device manufactured by Impedimed (Mansfield, Australia) can minimize both inter-rater and intra-rater variability seen with other measurement techniques. The device produces easily interpretable values measuring the changes in extracellular fluid when compared with the unaffected contralateral upper extremity. A score increase of 10 or more or one that registers outside the normal range should prompt intervention with a compression garment or a referral for physical therapy. Some investigators report that measurement of bioimpedance has improved sensitivity in detecting subclinical lymphedema up to 4 months prior to standard measurement changes. For these reasons, the current National Lymphedema Network (NLN) breast cancer–related lymphedema guidelines recommend that all breast cancer patients receive pre- and post-treatment measurements (of any type) on both arms and encourage the use of bioimpedance spectroscopy or infrared perometry as alternatives to tape measures, to limit measurement variations.
Is It All About the Nodes?
Over the years, many retrospective studies have reported risk factors for lymphedema. Axillary node dissection (ALND), mastectomy, obesity, radiation, infection, and ipsilateral upper extremity injury consistently rank as the most influential causes of damage or disruption to the lymphatic system and thus lymphedema. However, the magnitude of association between breast cancer treatment factors and lymphedema is inconsistent across studies. Tsai et al reviewed 98 studies conducted in the United States and Canada through January 2008 for lymphedema risk factors. The authors report a significantly increased risk for lymphedema in women undergoing mastectomy compared with lumpectomy (risk ratio [RR], 1.42; 95% confidence interval [CI], 1.15–1.76), ALND compared with no dissection (RR, 3.47; CI, 2.34–5.15), ALND compared with sentinel lymph node biopsy (SLNB) (RR, 3.07; CI, 2.20–4.29), radiation therapy vs no radiation therapy (RR, 1.92; CI, 1.61–2.28), and positive vs negative axillary lymph nodes (RR, 1.54; CI, 1.32–1.80). While this meta-analysis demonstrates a comprehensive, contemporary review of lymphedema risk factors, the heterogeneity of the data must be considered, as 11 different definitions for lymphedema were used and follow-up ranged between 1 month and 30 years.
Though the most intuitively obvious risk factor for lymphedema is the number of nodes removed, the relationship between the number of lymph nodes removed and the risk of lymphedema remains unresolved. Several retrospective studies have shown that the number of nodes removed and the risk of lymphedema do not correlate.[13,14] Others find an increasing risk of lymphedema as more nodes are removed.[15-17] Despite various lengths of follow-up, from 6 to 60 months, the prospective trials vetting SLNB as the standard of care for axillary staging demonstrate significantly reduced rates of lymphedema after SLNB (0 to 7%) compared with ALND (12% to 16%).[18-21] These studies certainly support the concept that the risk of lymphedema is proportional to the extent of axillary surgery. Importantly, though, they also confirm that even with SLNB a small but clear risk of lymphedema remains. A recent study by Goldberg et al suggests that it is not the number of lymph nodes removed but instead the degree of dissection and disruption of the lymphatic system that results in lymphedema. The authors reviewed 600 women having SLNB with a median follow-up of 5 years and found an overall incidence of lymphedema of 5%. When stratifying the data according to the number of nodes removed, they found no significant association between the mean, median, or range of number of nodes excised and lymphedema (P = .93). Furthermore, the authors completed a subset analysis of the women having more than 10 lymph nodes removed at SLNB. None of these women developed lymphedema. Interestingly, when these SLNB patients having > 10 nodes removed were compared with a separate group of women having 10 to 17 nodes removed at ALND, 11% of the ALND patients had measured lymphedema (P = .04). The fact that women having more than 10 nodes removed during SLNB did not develop lymphedema but women with the same number of nodes removed after ALND did reaffirms that the relationship between the nodes removed and lymphedema is complex. Perhaps it is the relative magnitude of lymphatic destruction and individual patient ability to form collateral lymphatic channels, rather than the number of nodes removed, that influences lymphedema risk. For example, a patient with many nodes removed at SLNB and no finding of lymphedema may have more lymphatic collaterals and therefore will have suffered relatively less lymphatic disruption despite a larger than “normal” number of SLNs removed. On the other hand, women having ALND and a relatively small number of total nodes excised may have suffered an overall greater degree of lymphatic disruption and therefore develop lymphedema. Unfortunately, the number of nodes within each patient’s nodal basin and the patient’s ability to protect or form new lymphatic collaterals during or after treatments is unknown. Therefore, simply the number of nodes removed may be insufficient to determine lymphedema risk.
If, however, the degree of lymphatic disruption or damage is the key driver of lymphedema risk, then it is plausible that radiation can also act primarily or synergistically to influence lymphedema risk. Axillary radiation alone is not without complications, as these patients can have a 2- to 4.5-fold increase in the risk of lymphedema. A recent meta-analysis by Shah and Vicini finds lymphedema in 9% to 65% of patients after lumpectomy alone (no nodal surgery) and regional nodal radiation and in 58% to 65% of women after mastectomy alone and regional nodal radiation. Additionally, the synergistic effect of surgery and radiation is well documented to result in a 3.5- to 10-fold higher risk of lymphedema when compared with surgery alone.[3,23,24]
Finally, accumulating data suggest that chemotherapy may also affect lymphatic destruction. Norman et al found anthracycline-based chemotherapy regimens increase the risk of lymphedema. Their prospective review of 631 breast cancer survivors followed for 5 years found an elevated hazard ratio (HR) of 1.46 (95% CI, 1.04–2.04) for lymphedema among breast cancer patients receiving anthracycline-based chemotherapy vs no chemotherapy. This risk persisted after stratifying for stage at diagnosis or number of positive nodes. Furthermore, the authors concluded that treatment combinations involving ALND or chemotherapy resulted in approximately four- to five-fold increases in the HRs for lymphedema (HR, 4.16 [95% CI, 1.32–12.45]for SLNB/chemotherapy/no radiation)compared with no treatment. While further validation of this finding is needed, the concept that chemotherapy independently influences lymphedema risk demonstrates that this risk can be affected not only by locally directed therapies but also by systemic ones.
Changes in Surgical Management
Changes in the surgical management of the axilla have been instrumental in reducing axillary morbidity, especially lymphedema. SLNB is now the standard of care for axillary staging, as its accuracy, low false-negative rate, and low rate of axillary recurrence have been documented in more than 70 studies. Current research to further minimize morbidity from axillary surgery focuses on reducing the need for completion axillary dissection in the setting of a positive axillary node. The American College of Surgeons (ACOSOG) Z0011 trial published in early 2011 documents the first prospective randomized data showing that at a median of 6 years follow-up, SLNB alone does not result in inferior survival in women with T1 or T2 tumors and one or two positive sentinel nodes who received breast conservation surgery, chemotherapy and/or hormonal therapy, and whole breast radiation therapy. In the study, more than 70% of the women with a positive SLNB had no additional positive nodes, suggesting that in select cases SLNB can be both diagnostic and therapeutic. While this trial is viewed as practice-changing, it is important to note that the trial called for standard whole breast radiation but did not standardize the radiation tangents other than to say that third field–directed axillary radiation was forbidden. As a result it is possible that radiation tangents were adjusted higher in node-positive women who were randomized to the SLNB-only arm. If standard radiation tangents cover approximately 30% to 50% of the level I and 25% of the level II axillary nodes, then adjusting the tangents “higher” has the intention of covering more nodes than standard. These data are important from the perspective of morbidity, as an adjustment in the tangents may result in increased rates of axillary morbidity above what is seen with standard whole breast tangents and SLNB alone. The radiation ports are currently being evaluated retrospectively. An additional prospective trial, ACOSOG Z1071, is evaluating the validity, accuracy, and false-negative rate of SLNB after neoadjuvant chemotherapy in women who present with node-positive disease at diagnosis. The investigators hypothesize that SLNB after neoadjuvant chemotherapy will appropriately and reliably stage the axilla and therefore allow elimination of ALND in women converted by neoadjuvant chemotherapy from node-positive to node-negative status. The trial accrued quickly and results are pending. Remaining questions to be answered include that of which women may be able to forgo axillary surgery (SLNB or ALND) altogether.