Morbidity after conventional dissection of axillary lymph nodes in breast cancer patients

Axillary lymph node dissection (ALND), although controversial in specific situations, remains an integral part of surgical treatments in patients with invasive breast cancer and axillary lymph node metastases [1]. Specifically, this treatment is applicable in patients with tumors that are considered N1 or N2 according to the TNM staging system [2]. ALND was replaced in clinical practice by sentinel lymph node biopsy (SLNB) in patients that lack axillary lymph node involvement (N0) and some N1 patients [3], due to its reduced morbidity [4‐7]. However, despite widespread mammography use for disease screening and early diagnosis, approximately one-third of patients in the U.S. suffer from tumors that have spread to the regional lymph nodes at diagnosis according to the Surveillance, Epidemiology and End Results (SEER) database [8]. In Brazil, according to National Department of Health statistics (2012), only 18% of these cancers are confined to the breast at diagnosis (pathologic stage), despite efforts to provide screening mammography for all women aged older than 40 years [9]. This late diagnosis means that, proportionally, more patients are submitted to ALND in Brazil than their American counterparts.

Currently, conventional ALND involves lymph node resection in levels I and II, as described by Berg [10]. Such dissections serve therapeutic functions and enable disease stage and prognosis assessments [11]. Unfortunately, ALND is primarily responsible for functional surgical treatment sequelae, including lymphedema, paresthesia, range-of-motion restriction, and pain in the arm ipsilateral to the lymph node dissection. Although esthetic sequelae that are caused by partial or total surgical breast resection can be reversed or minimized by reconstructive surgery methods that include prosthetics and tissue flaps, little can be done to correct the functional sequelae [12].

Since Halsted advocated the ‘systematic cleaning out of the axilla’ as an essential part of the operation ‘for the cure of cancer of the breast’ in 1907 [13], efforts to reduce the radicality and extent of tissue resection in ALND have been proposed. In 1948, Patey and Dyson [14] initially preserved the pectoralis major muscle but resected the pectoralis minor muscle along with ALND. Subsequently, Auschincloss [15] and Madden [16] proposed and closely described ALND with preservation of the pectoralis major and minor muscles in 1963 and 1965, respectively. The dissection of Berg level III has ceased in recent years [17], and the number of lymph nodes resected have decreased. A review of 21,992 women in the California Cancer Registry, who underwent ALND between 2004 and 2008, revealed an average of 11.4 ± 7.4 dissected lymph nodes per ALND [18]. In the 1960s, Auschincloss described an average of 38 lymph nodes per ALND [15]; this difference suggests that ALND has become increasingly conservative. Previous studies have shown that the incidence rates of complications and sequelae in the arm, including lymphedema, are directly related to the locoregional treatment radicality, which involves surgery and radiation therapy [19, 20]. In the 1990s, studies showed that the incidence of lymphedema was decreasing due to more conservative approaches to the axilla [21].

Accordingly, this study aims to assess the current prevalence of ALND-associated complications in breast cancer patients.

The mean interval between ALND and analysis was 35 months (standard deviation, 18 months) with a range of 7 to 60 months. The average dissected lymph node number per patient was 14 ± 4 (range, 6-30 lymph nodes). None of the patients exhibited axillary recurrence during the analysis period. The patient and treatment characteristics are shown in Table 1.

Lymphedema was the least frequent complication and was found in only 3.5% (n = 7) of patients. Paresthesia was the most frequent complication and was observed in 53% (n = 106) of patients. Arm range-of-motion restriction, which was observed in 24% (n = 48) of the patients, and pain, which was reported by 27.5% (n = 55) of the interviewed patients, occurred with moderate frequency. Because 95% (n = 190) of the evaluated patients received breast radiotherapy and 96% (n = 192) received some type of chemotherapy to treat their breast cancer, both of these treatment modalities were excluded as potential variables from statistical analysis.

ALND’s importance in the staging, prognostic assessment and local control of breast cancer has long been proven [23]. Local control of breast cancer also results in longer survival [24], and failing to ‘clean’ the axilla could mean a loss of disease control with distant metastases, such as that noted in the NSABP B-04 study [25]. If not for the morbidity, the need for ALND would not be debated in cases of SLNB positivity [3]. In contrast, the incorporation of SLNB in breast cancer treatment has benefited many women with the disease at an early stage, confined to the breast, avoiding ALND. This treatment course substantially decreased functional sequelae risks in the arm that result from surgical armpit manipulation. Because women are increasingly surviving breast cancer [26], ALND morbidity is generating a high social, psychological, and financial cost.

Unfortunately, 40.4% of patients in Brazil already present with stage III and IV disease at diagnosis [9]. In this study, 51.5% of the patients (n = 103) had metastases in the axillary lymph nodes. The N0 patients who underwent ALND (48.5%, n = 97) had contraindications to SLNB (for example, inflammatory breast cancer, T4 tumors, clinically positive axilla) or refused the procedure.

Currently, conservative surgery combined with breast radiation therapy is considered as effective as total mastectomy for the local control of breast cancer [27, 28]. This combination treatment was administered to 82.5% (n = 165) of the patients in this study. Previous studies have shown that the type of treatment used influences the morbidity prevalence. Schünemann and Willich [19] evaluated 5,868 patients with breast cancer treated from 1972 to 1995 and demonstrated that the addition of radiotherapy to modified radical mastectomy increased the lymphedema incidence from 19.1% to 28.9%. DiSipio et al. [20], in a systematic review and meta-analysis that evaluated 72 studies, associated chemotherapy with lymphedema. These two variables (radiotherapy and chemotherapy) were not evaluated in this study, because the vast majority of the patient cohort (95% and 96% respectively) was subjected to these therapeutic modalities.

Lymphedema is by far the most serious and difficult to treat complication that has the greatest effect on a patient’s quality of life. However, lymphedema is not the only ALND-related complication. The literature primarily reports lymphedema, paresthesia, pain, and range-of-motion restriction as complications of ALND. The risk of complications correlates positively with the radical nature of ALND [19, 29].

The lymphedema incidence and prevalence described in the literature vary widely, possibly due to different measurement methods and intervals between ALND and lymphedema measurement. DiSipio et al.[20] observed a 21.4% lymphedema incidence after analyzing 30 prospective cohort studies. Therefore, the 3.5% prevalence observed in this study is below the literature average. The lymphedema measurement method used in this study (circumference measurement) is simple, easily reproducible, and the most used approach in the medical literature. Using this method, the average lymphedema incidence observed was 14.8%, as described by DiSipio et al.[20] Another study using the same lymphedema assessment criteria (upper limb measurement difference of >2 cm between the arms) was conducted in Greece by Keramopoulos et al.[30] and reported a 17% lymphedema incidence. Possible explanations for this unexpectedly low lymphedema prevalence are the lack of systematic dissection of level III, the lack of interpectoral space exploration (Rotter lymph nodes), and the attempted preservation of the pectoral muscles. Evidently, this approach results in fewer dissected lymph nodes, as noted in the present study with an average of 14 dissected lymph nodes per patient. However, none of the patients showed axillary recurrence during the analysis period. Although this number is lower than the 38 dissected lymph nodes per patient reported in the 1960s [15], 10 dissected lymph nodes per patient is currently considered to be an adequate minimum number [31, 32]. It is also important to note that the time between ALND and analysis was less than 2 years in some patients. The review conducted by DiSipio et al. [20] demonstrated that lymphedema appears to increase 2 years after diagnosis of or surgery for breast cancer. The fact that no patient had the axillary nodal region included in the radiation field may also have contributed to reduce the risk of lymphedema.

Finally, another important aspect to be considered is that difference of >2 cm between the arms may be too much to objectively evaluate lymphedema morbidity and an investigation with upper limb lymphoscintigraphy, for example, would have helped to detect lymphatic impairment much earlier than clinically. Therefore, this may have led to an underestimation of lymphedema prevalence in our results.

Paresthesia was the most frequent complication in this study and was found in more than half of the patients (53%, n = 106). Veronesi and colleagues [22] reported a prevalence of 68%, and Warmuth et al.[12] reported a prevalence of 35%. This finding is related to the intercostobrachial nerve section that crosses the axilla and is transected during ALND. However, paresthesia is a minor complication that neither results in complaints nor limits the quality of life in most cases [33]. There was no intent to preserve the intercostobrachial nerve (ICBN) during ALND for the patients evaluated in the present study.

Pain and range-of-motion restriction occurred in 27.5% (n = 55) and 24% (n = 48) of the patients, respectively. The range-of-motion restriction was evaluated according to the degree of arm abduction, a method that has been previously used [34]. The pain and range-of-motion restriction incidence rates vary widely in literature reports. Warmuth et al.[12] identified an 8% limitation in arm movement and a 30% incidence of pain in 432 patients 2 to 5 years after ALND. Kootstra et al.[35] evaluated 76 women and observed that 70% had clinical relevant impairments in the shoulder and arm 7 years after ALND. It is noteworthy that 62.5% (n = 30) of patients who presented with range-of-motion restriction (n = 48) had only mild range-of-motion restriction (120° to 179° abduction). The range-of-motion restriction might be overestimated in the present study because a rangeofmotion of at least 160° can be considered normal [36].

This study also found a correlation between lymphedema and single-incision surgery. This fact was not addressed in related previous studies. A possible explanation is that in breast-conserving surgery, attempts to approach the breast (quadrantectomy) and axilla (ALND) through the same incision sometimes create tunnels that complicate the correct identification of planes that involve the axilla and the preservation of the periaxillary subcutaneous tissue.

The present study also found an association between paresthesia and advanced patient age. Ververs and colleagues reported an opposite finding [37]. They discovered an inverse relationship between arm paresthesia and patient age.

The present study showed that pain, a subjective datum, was associated with surgery ipsilateral to the dominant arm. This association was also found in a previous study by Hayes and colleagues [38].

In conclusion, although SLNB has benefited many women with breast cancer, many patients still require ALND, despite its associated risk for morbidity. The results of this study showed that the lymphedema prevalence, the most undesirable ALND side effect, was low in relation to other evaluated symptoms. Further, larger prospective studies are required to fully assess the exact morbidity incidence following ALND.