Lymphedema after sentinel lymphadenectomy for breast carcinoma

Authors

  • Stephen F. Sener M.D.,

    Corresponding author
    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Surgery, Northwestern University Medical School, Chicago, Illinois
    • Evanston Northwestern Healthcare, 2650 Ridge, Burch 106, Evanston, IL 60201
    Search for more papers by this author
    • Fax: (847) 570-1330

  • David J. Winchester M.D.,

    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Surgery, Northwestern University Medical School, Chicago, Illinois
    Search for more papers by this author
  • Carole H. Martz R.N., M.S.N.,

    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Surgery, Northwestern University Medical School, Chicago, Illinois
    Search for more papers by this author
  • Joseph L. Feldman M.D.,

    1. Department of Medicine, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Medicine, Northwestern University Medical School, Chicago, Illinois
    Search for more papers by this author
  • Jean A. Cavanaugh M.D.,

    1. Department of Medicine, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Medicine, Northwestern University Medical School, Chicago, Illinois
    Search for more papers by this author
  • David P. Winchester M.D.,

    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    2. Department of Surgery, Northwestern University Medical School, Chicago, Illinois
    Search for more papers by this author
  • Beth Weigel R.N., B.S.N.,

    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    Search for more papers by this author
  • Kathleen Bonnefoi R.N., M.S.,

    1. Department of Surgery, Evanston Northwestern Healthcare, Evanston, Illinois
    Search for more papers by this author
  • Katina Kirby M.S.,

    1. Department of Medicine, Evanston Northwestern Healthcare, Evanston, Illinois
    Search for more papers by this author
  • Claudia Morehead P.T.

    1. Department of Medicine, Evanston Northwestern Healthcare, Evanston, Illinois
    Search for more papers by this author

Abstract

BACKGROUND

Initial studies of sentinel lymphadenectomy for patients with breast carcinoma confirmed that the status of the sentinel lymph nodes was an accurate predictor of the presence of metastatic disease in the axillary lymph nodes. Sentinel lymphadenectomy, as an axillary staging procedure, has risks of morbidity that have yet to be defined.

METHODS

Patients were enrolled in a two-phase protocol that included concurrent data collection of patient characteristics and treatment variables. During the first (validation) phase, 72 patients underwent sentinel lymph node excision followed by a level I–II axillary dissection. After the technique had been established, the second phase commenced, during which only patients with positive sentinel lymph nodes underwent an axillary dissection.

RESULTS

During the second phase, lymphedema was identified in 9 of 303 patients (3.0%) who underwent sentinel lymphadenectomy alone and in 20 of 117 patients (17.1%) who underwent sentinel lymphadenectomy combined with axillary dissection (P < 0.0001). Of 303 patients who underwent sentinel lymphadenectomy alone, 8 of 155 patients (5.1%) with tumors located in the upper outer quadrant and 1 of 148 patients (0.7%) with tumors in other locations developed lymphedema (P = 0.012).

CONCLUSIONS

The risk of developing lymphedema after undergoing sentinel lymphadenectomy was measurable but significantly lower than after undergoing axillary dissection. Tumor location in the upper outer quadrant and postoperative trauma and/or infection were identifiable risk factors for lymphedema. Cancer 2001;92:748–52. © 2001 American Cancer Society.

Prior to the development of sentinel lymphadenectomy for patients with breast carcinoma, it was common practice to completely remove level I–II axillary nodes to provide accurate staging and potential therapeutic value. Only 25–35% of patients had metastatic lymph nodes, yet all accepted the possibility of complications after undergoing axillary dissection. Initial studies of sentinel lymphadenectomy for patients with breast carcinoma confirmed that the status of the sentinel lymph node(s) was an accurate predictor of the presence of metastatic disease in the axillary lymph nodes.1–7 Thus, patients with negative sentinel lymph nodes were spared the morbidity of axillary dissection.

However, as an axillary staging procedure, sentinel lymphadenectomy also has potential risks of morbidity that have yet to be defined. In this report, the authors assessed the risk of lymphedema in a consecutive series of patients with breast carcinoma who were treated with sentinel lymphadenectomy.

MATERIALS AND METHODS

Sentinel lymphadenectomy for patients with carcinoma of the breast was initiated by three surgeons (S.F.S., D.J.W., and D.P.W.) in late 1996 as a pilot validation study at the authors' institution.7 After informed consent was obtained, patients were enrolled in a protocol approved by the Institutional Review Board. Data regarding patient characteristics and treatment variables were collected concurrently in a computerized data base.

During the validation phase (first phase) of the study, 72 patients underwent sentinel lymphadenectomy followed by level I–II axillary dissection. Modifications of the technique were made during this phase to optimize migration kinetics of the injected sulphur colloid. Subsequent to the validation phase, the second phase commenced, during which only patients with positive sentinel lymph nodes underwent level I–II axillary dissection. All patients received injections of 37 mBg (1 mCi) technetium 99m sulphur colloid, which was filtered through a 0.22-μm filter and diluted in an 8-cc volume. The injection was delivered in a four-quadrant, peritumoral distribution. All patients had preoperative lymphoscintigraphy. The elapsed time between injection and sentinel lymph node excision was 16–20 hours. Patients with nonpalpable tumors underwent wire localization before colloid injection, and a repeat mammogram on the day of surgery confirmed the correct guide-wire position.

Intraoperative, transcutaneous identification of sentinel lymph nodes was accomplished with a hand-held gamma probe. Radiolabeled lymph nodes were examined off the operative field. If ex vivo counts per second exceeded three times the lymph node basin background level, then the lymph node was considered to be a sentinel lymph node.

For patients who underwent lumpectomy as the definitive surgical procedure, radiotherapy was delivered to the remaining breast in 33 fractions. A 6-MeV linear accelerator was used to treat the whole breast to 5040 centigrays (cGy). Electrons (12 MeV) were then used to boost the lumpectomy site with an additional 1000 cGy. With the exception of Patient 6 (see Table 2), there was no attempt to include the axilla in the radiation port.

Beginning with the second phase of the study (late 1997), information related to lymphedema was collected. Included were arm measurements that were obtained before surgical procedures and at intervals during follow-up. Patients were taught standard lymphedema precautions and were trained to perform range-of-motion exercises, regardless of whether they had undergone a level I–II axillary dissection. An aggressive position was established toward the evaluation and treatment of symptoms related to lymphedema, as self-reported by patients between and during follow-up visits. Patients were reexamined at 6-month intervals after undergoing definitive treatment for breast carcinoma.

The severity of lymphedema was classified according to a modification the original definitions of Stillwell.8 Postoperative volume differences < 20% between arms were defined as mild, differences of 21–40% were defined as moderate, and differences > 40% were defined as severe.

Factors associated with lymphedema were defined as traumatic if the physician, physical/occupational therapist, nurse clinician, and patient agreed that an identifiable injury or event was proximate in time to the development of lymphedema. Events that were included in the operational definition of trauma were insect bites, intercontinental air travel, the use of venipuncture and blood pressure cuff on the operated arm, lifting groceries, and vigorous athletic activity early in the postoperative period.

Statistical analyses were performed using the chi-square test. Differences were accepted as significant if P values were < 0.05.

RESULTS

From December 1996 through April 2000, 492 consecutive patients with invasive breast carcinoma underwent sentinel lymphadenectomy. There were 72 patients included in the validation (first) phase, in which sentinel lymphadenectomy was followed routinely by level I–II axillary dissection. Of the 420 patients who were treated during the second phase of the study, 117 patients (28%) had metastatic disease in sentinel lymph nodes. These patients underwent level I–II axillary dissection. Thus, there were 303 patients who underwent sentinel lymphadenectomy alone as the axillary staging procedure. At the time of this analysis, the median follow-up was 24 months for the entire cohort and 19 months for patients who underwent sentinel lymphadenectomy alone.

Incidence of Lymphedema

Lymphedema developed in 5 of 72 patients (6.9%) who underwent sentinel lymphadenectomy followed by obligatory axillary dissection during the first phase. All five patients were lymph node positive. However, data regarding lymphedema were not collected consistently during that period.

In the second phase, lymphedema developed in 9 of 303 patients (3.0%) after undergoing sentinel lymphadenectomy alone (lymph node negative patients) and in 20 of 117 patients (17.1%) after undergoing sentinel lymphadenectomy and axillary dissection (lymph node positive patients) (Table 1). Thus, the risk of lymphedema was finite and was significantly lower for patients who underwent sentinel lymphadenectomy alone compared with patients who underwent sentinel lymphadenectomy with axillary dissection (P < 0.0001).

Table 1. Development of Lymphedema in 303 Patients who Underwent Sentinel Lymphadenectomy Alone and in 117 Patients who Underwent Sentinel Lymphadenectomy Followed by Axillary Dissection
LymphedemaSentinel lymphadenectomy aloneSentinel lymphadenectomy followed by axillary dissection
Tumor in UOQOther locationsTotalTumor in UOQOther locationsTotal
  1. UOQ: upper outer quadrant.

Positive81912820
Negative147147294504797
Total1551483036255117

Tumor Location and Risk of Lymphedema

During the second phase, 117 patients underwent sentinel lymphadenectomy followed by axillary dissection (lymph node positive patients). Of these, lymphedema developed in 12 of 62 patients (19.3%) with tumors located in the upper outer quadrant and in 8 of 55 patients (14.5%) with tumors in other locations (Table 1). Thus, for lymph node positive patients who required axillary dissection, tumor location in the upper outer quadrant was not associated with increased risk of lymphedema (P value not significant).

Also during the second phase, 303 patients underwent sentinel lymphadenectomy alone (lymph node negative patients). Of these, lymphedema developed in 8 of 155 patients (5.1%) with tumors located in the upper outer quadrant and in 1 of 148 patients (0.7%) with tumors in other locations (Table 1). Thus, for lymph node negative patients who underwent sentinel lymphadenectomy alone, tumor location in the upper outer quadrant was associated with increased risk of lymphedema (P = 0.012).

Characteristics of Lymphedema after Sentinel Lymphadenectomy Alone

The median time from the surgical procedure to the clinical recognition of lymphedema was 4.0 months (range, 1.5–16.0 months; mean, 4.3 months) for nine patients who underwent sentinel lymphadenectomy alone. Edema was characterized as mild in eight patients and moderate in a ninth patient (Patient 2). (Table 2.) In three patients (Patients 3, 6, and 9), clinical symptoms related to intercostobrachial cutaneous neuropathy were prominent features of the presentation. The edema in Patient 2 may have been related to a weight gain of > 10 pounds after undergoing mastectomy and sentinel lymphadenectomy and to the development of a postoperative wound infection. She had a discrepancy in water displacement volume of 450 cc between the operated and unoperated arms.

Table 2. Total Number of Lymph Nodes Removed, Volumetric Differences, and Severity of Edema in Nine Patients who Developed Lymphedema After Undergoing Sentinel Lymphadenectomy Alone
Patient no.Total no. of lymph nodes removedVolume difference (cc)aEdemab
  • a

    Values indicate the postoperative water displacement volume difference between arms.

  • b

    The location of edema is indicated, with its severity shown in parentheses.

  • c

    Not measured.

  • d

    Unmeasurable shoulder edema.

12175Arm (mild)
23450Arm (moderate)
33cShoulder (mild)
44135Arm (mild)
54180Arm, breast (mild)
63−10Shoulder (mild)d
7375Arm (mild)
8545Arm (mild)
9525Shoulder (mild)

Discernible risk factors for lymphedema in the nine patients are shown in Table 3. Five of the nine patients underwent a total mastectomy, and four patients underwent lumpectomy followed by radiotherapy to the breast. In patients who developed lymphedema, the median number of lymph nodes removed was 3 (range, 2–5 lymph nodes), and the median tumor size was 18 mm (mean, 17.6 mm). These values were similar to those for the 294 patients who underwent sentinel lymphadenectomy alone and did not develop lymphedema.

Table 3. Identifiable Risk Factors in Nine Patients who Developed Lymphedema After Undergoing Sentinel Lymphadenectomy Alone for Breast Carcinoma
FactorNo. of patients
Tumor location in upper outer quadrant8
Breast radiotherapy4
> 1 Breast operation3
> Ten-pound weight gain after operation3
Trauma to arm, including wound healing complications (infection, etc.)7
Axillary radiotherapy1

Tumor location in the upper outer quadrant was combined with infection in three patients and with trauma in four patients. In two patients with tumors in the upper outer quadrant, intercontinental air travel was associated temporally with the onset of lymphedema. In Patient 6 (Table 2), in whom neither trauma nor infection was identified, factors contributing to edema may have been tumor location in the upper outer quadrant and adjuvant treatment. The therapy consisted of doxorubicin, cyclophosphamide, paclitaxel, and radiotherapy to the breast and axilla. This patient had a micrometastasis in one lymph node, leading to the recommendation for axillary radiation. Patient 9 (Table 2) developed mild axillary and chest wall edema subsequent to becoming pregnant 4 months after undergoing mastectomy.

DISCUSSION

Incidence of Lymphedema after Level I–II or Complete Axillary Dissection

In a review from 1998, Petrek and Heelan9 performed a comparative analysis of lymphedema incidence in seven studies that were published in the 1990s.10–16 Patients underwent different surgical and radiotherapeutic procedures for breast carcinoma in the United States, England, Italy, France, and Germany. In the seven retrospective studies, the incidence rate of objective lymphedema ranged from 6%11 to 30%.14 The study populations were diverse, and arm circumferences or volumetric equipment were used to measure edema at various time intervals after axillary dissection. Comparisons between studies were made difficult by the lack of descriptions of the breast carcinoma treatments employed in the study populations and the inability, in all seven reports, to define the total number of patients at risk for developing edema in a particular population (the denominator).

In the same review, Petrek and Heelan9 reported an analysis of 272 patients who were diagnosed with breast carcinoma from 1976 through 1978 at Memorial Sloan-Kettering Cancer Center. Enlargement ≥ 2 inches in circumference over the contralateral arm was found in 32 patients (12%), and enlargement < 2 inches in circumference was found in 43 patients (16%). In addition to the 75 patients (28%) with measurable edema, 47 patients (17%) stated that the arm felt swollen, but there was no measurable difference in arm circumference.

Risk Factors for the Development of Lymphedema

Surgical procedures and radiotherapy in the axilla contribute to the development of arm lymphedema. The risk and severity of edema have been correlated with the extent of axillary surgery.17, 18 The use of axillary dissection or axillary radiotherapy, but not both, has been associated with a risk of lymphedema as great as 27%, with a trend toward higher risk with complete axillary clearance of lymph nodes.16, 19–22 The highest rates of lymphedema, ranging as high as 44%, have been reported in patients who were treated with complete axillary dissection plus axillary radiotherapy.16

Other factors that have been reported to increase the risk of arm edema after radiotherapy are advanced disease stage at diagnosis (especially lymph node),16 obesity,10 prolonged presence of seroma and delayed wound healing,23, 24 wound infection, and cellulitis of the operated arm. Older age and the use of adjuvant chemotherapy have been reported in some, but not all, studies to increase the complication rates of radiotherapy.20, 25 In addition, data from the current study demonstrated that tumor location in the upper outer quadrant was associated with an increase in the risk of lymphedema.

Incidence of Lymphedema after Sentinel Lymphadenectomy

Schrenk et al.26 reported a prospective, nonrandomized study of 35 patients after undergoing sentinel lymphadenectomy and 35 patients after undergoing level I–II axillary dissection. Postoperative follow-up was approximately 16 months in both groups. There was no significant difference in preoperative versus postoperative measurements of the upper arm and forearm in the sentinel lymphadenectomy group, whereas a significant increase was identified in the greatest dimension of the upper arm and forearm in the axillary dissection group. Subjective lymphedema was not reported in any of the 35 patients after sentinel lymphadenectomy. The authors concluded that the postoperative morbidity associated with sentinel lymphadenectomy was negligible compared with that of axillary dissection.

However, with a larger study sample size, the current analysis demonstrated that the risk of lymphedema after sentinel lymphadenectomy was measurable but significantly lower than that after sentinel lymphadenectomy combined with axillary dissection. Because of the complexities of the interrelations between neuropathy and fluid accumulation in the arm, breast, and chest wall, it was not possible to compare the severity of lymphedema for the two patient subsets.

Lymphedema after therapy for breast carcinoma may be acute or chronic. Acute edema occurs shortly after the surgical procedure and commonly is transient. Chronic lymphedema may develop months or years after primary treatment for malignant disease and is more resistant to therapy. Because the nine patients in this series with lymphedema after sentinel lymphadenectomy presented within a relatively short median follow-up, they should be regarded as having the acute form of lymphedema. Longer follow-up will better define the evolution to chronic forms of lymphedema in this patient cohort.

The tumor location, the extent of axillary staging, and the presence of trauma and/or infection have been identified in this study as factors that were associated temporally with the onset of lymphedema. However, the distinction should be made between causative factors and associated factors. Although they were associated temporally with the onset of symptoms, the clinical observations contained herein do not establish a true cause-and-effect correlation between these factors and the development of lymphedema.

Ancillary