Lymphedema is caused by a build-up of lymph fluid in tissues after breast cancer (BC) surgery and radiotherapy. It is one of the most common complications after BC surgery, and approximately 15% to 46% of patients with BC suffer this complication after BC treatment.1–7 Lymphedema can result not only in physical discomfort and disfigurement but also in substantial impairment of daily activities. As a cure has not been established at this time, prevention of lymphedema is of key importance. To aid in educating patients about the prevention of lymphedema we should know what, if any, factors contribute to an increased risk of lymphedema, and which subgroups of patients are at an increased risk. Once the factors that influence the development of lymphedema are clarified, such findings can be used to develop preventive measures.
In general, risk factors of lymphedema are classified into 3 categories that are as follows: treatment-related factors, disease-related factors, and patient related factors. The type of surgery, radiotherapy, chemotherapy, and other combined treatments are treatment related factors for lymphedema.8,9 Disease-related factors for lymphedema include tumor stage, the number of lymph nodes that are excised, and the location of tumor.8–10 Patient-related factors that have been associated with lymphedema include age at lymphedema diagnosis, body mass index (BMI), infection of surgical side upper extremity, and limb overuse.8–10
The risk factors of lymphedema in patients with BC have been studied in several trials but the etiology of lymphedema is still not completely understood. A predicting tool for lymphedema should be created to guide physicians and patients with BC to understand how to prevent and control lymphedema. Our group tested a set of risk factors of lymphedema and found that postoperative infection of surgey side upper extremity, level of hand use, and BMI were 3 statistically significant factors to predict the risk and severity of lymphedema.9
The aim of our study is to investigate whether combinations of these 3 risk factors could accurately predict lymphedema, and to estimate the probability of lymphedema development in the final model. The results can be useful to guide healthcare professionals and patients with BC to prevent or decrease? better maybe?? the risk of lymphedema in favor of determining the most powerful predictors of lymphedema for the general population.
MATERIALS AND METHODS
A total of 2983 female patients having breast/axillary surgery were recruited at Magee-Womens Hospital of the University of Pittsburgh Medical Center between 1990 and 2000. However, only 51 patients with lymphedema had adequate data for analysis (the severity of lymphedema). The design was a n:m matched case-control study; data were analyzed on 51 patients with LE and 126 available controls were matched on age, radiation therapy, and type of operation.
As n:m matching was carried out, there were varying numbers of cases and controls in the matched sets. A total of 177 patients (51 cases and 126 controls) were matched on age (<45, 46 to 54, 55 to 64, >65 y), radiation therapy (yes/no), and type of operation (segmental mastectomy, modified radical mastectomy, and modified radical mastectomy with reconstruction), and were categorized into n:m matched sets. The SAS System (SAS Institute Inc, Cary, NC) version was used for analysis.
Three risk factors, BMI (<25 kg/m2, ≥25 kg/m2), postoperative ipsilateral upper extremity infection (yes/no), and occupational/hobby hand use (low, medium-high) formed 8 combinations. Since 1 combination included no cases, the probability of developing lymphedema was estimated for 7 combinations (Table 2). The Bayes' theorem was used to estimate the probability of developing LE using (1) published estimates of LE incidence rates and (2) the estimated proportions of risk factor combinations in cases and controls obtained from our patient sample. Different estimations for LE incidence rates were selected from published papers, and the sensitivity of lymphedema probabilities based on these were discussed.
Table 1 shows the comparison of LE incidence among several published papers. LE probabilities of 7 risk factor combinations for 6 different published incidence rates of LE were calculated (Table 2). The incidences ranged from 16% to 46.3% depending on the definition of LE, the method of measurement, the length of follow-up, and choice of therapy in each independent study. Prediction of LE was higher in patients who had the defined combinations stated as set 4, 7, and 8 in Table 2. The highest estimated LE probability was 76.7% for BC patients with BMI≥25 kg/m2, infection, and medium/high of hand use when the LE incidence rate of 16% was used. We could not calculate the estimation for the set 3 since we had no patients qualifying these criteria. (BMI<25, infection, low level hand use).
Figure 1 presents the estimated LE probabilities by comparing the different LE incidence rates in each risk factor combination. The lines for combinations 1 and 5 are nearly superimposed. The difference between these 2 risk factor combinations is the BMI category. Although the estimated probabilities in all combinations increased from 16% to 46.3%, it is important to assess how sensitive the estimated probabilities at different LE incidence rates.
Figure 2 presents the percentage change of the estimated LE probabilities by comparing different LE incidence rates in each risk factor combination. When an incidence jumped to 46.3%, a greater percentage increase was found, especially in combinations 1 and 5. Higher estimated LE probabilities correlated with a less percentage change in different incidence rates. Patients in combination 8 had a much higher LE probability than patients in other combinations, and combination 8 also showed less variability because it had a flat and smooth line of depicting each percentage difference of LE probabilities between 2 incidence values. Overall, the estimated probabilities in combination 8 were relatively insensitive to change in incidence values between 17% and 25.5%.
In general, LE does not occur in all patients with BC patients treated in similar manner, but the risk of LE development does increase as a result of predisposing factors, such as obesity and infection. There are no definite prediction rules or algorithm to assess LE development and it is not to be expected that any specific intervention could be adapted to all patients with BC. However, a general predictive model for LE still can be built to estimate the risk of LE. For instance, a predicting tool (nomogram) to determine the necessity of complete ALND for patients with BC with sentinel lymph node metastases has been developed and validated by different medical centers. A similar prospective predicting tool for lymphedema can be developed along the same lines as the sentinel lymph node-predicting nomogram.
The purpose of this study was to estimate LE probability after BC surgery by 3 confirmed risk factors and to assess its variability in predicting LE. Although some other risk factors associated with LE development have been discovered,8,9,11–13 our earlier findings indicated that patient with BMI ≥25 kg/m2, infection of surgery side upper extremity, and medium/high level occupational/hobby hand use have significantly higher high risk of LE after BC surgery (Table 3). Despite the disease or treatment related factors such as tumor size, LN involvement, LN dissection, RT are not modifiable risk factors, BMI, level of hand use and infection are subject to prevention or modification. After we matched age, radiation, and surgical treatment, the predicted probability of LE varied from 77% to 94% with these 3 risk factors, depending on which incidence rate was used (Table 1).
Swelling may occur at any point after axillary node dissection and radiation therapy, beginning immediately after or even delayed by several years.11 In the literature, a broad range of incidence rates of LE varies widely from 6% to 70% after BC surgery.14–16 In this study, 16% to 46.3% incidence rates were used for the assessment of LE probabilities. Our study showed LE probabilities estimated between 19.5% and 20.7% of LE incidence being similar for each risk factor combination (Table 1).
We found that the predicted LE probabilities for patients in combination 8 were insensitive among different LE incidence rates. These predicted probabilities did not have a significant difference along increased incidence rates. However, the predicted LE probabilities in combinations 1 and 5 were sensitive to incidence rates. Especially when LE incidence changed to 46.3%, the estimated probabilities in these 2 combinations increased tremendously. We could conclude that, the combination high BMI >25, infection and high/medium level of hand use comprise of a considerably steady risk of LE development compared to other risk factor combinations regardless of the LE incidence used.
The importance of this study is to determine what, if any, factors contribute to an increased risk of LE as well. As to establish which subgroups of patients are at an increased risk, once the factors that influence the development of LE are clarified, such findings can be used to develop preventive measures.
This model represents a significant improvement over estimates based on 3 risk factors, but it is limited by the small number of patients. The model requires to be tested on large group of patients. Another limitation is the fact that the controls were sampled to match cases. Therefore, it could not be considered as a random sample control population.
This study shows that patients with low incidence for LE are more prone to develop LE if the predisposing factors are controlled poorly compared to the high incidence patients whom the predisposing factors are avoided. A well-established lymphedema-predicting tool for BC survivors should be taken into consideration in the future.
1. Petrec JA, Pressman PI, Smith RA. Lymphedema
: current issues in research and management. CA Cancer J Clin. 2000;50:292–307
2. Kissin MW, Querci della Rovere G, Easton D, et al. Risk of lymphedema
following the treatment of breast cancer
. Br J Surg. 1986;73:580–584
3. Werner RS, Mc Cormic B, Petrek J, et al. Arm edema in conservatively managed breast cancer
: obesiti is a major predictive factor. Radiology. 1991;180:177–184
4. Lin PP, Alison DC, Wainstock J, et al. Impact of axillary lymph node dissection on the therapy of breast cancer
patients. J Clin Oncol. 1993;11:1536–1544
5. Keramapoulus A, Tsionou C, Minaretzis D, et al. Arm morbidity following treatment of breast cancer
with total axillary dissection: a multivariated approach. Oncology. 1993;50:445–449
6. Deutsch M, Flickinger C. Arm edema after lumpectomy and breast irradiation. Am J Clin Oncol. 2003;26:229–231
7. Clark B, Sitzia J, Harlow W. Incidence
and risk of arn oedema following treatment for breast cancer
: a three-year follow-up study. QJM. 2005;98:343–348
8. Galler BM, Vacek PM, O'Brain P, et al. Factors associated with arm swelling after breast cancer
surgery. J Women's Health. 2003;12:921–930
9. Soran A, D'Angelo G, Begoviç M, et al. Breast cancer
: What are the significant predictors and how they affect the severity of lymphedema
? Breast J. 2006;6:536–543
10. Coen JJ, Taghian AG, Kachnic LA, et al. Risk of lympedema after regional nodal irradiation with breast conservation therapy. Int J Radiat Oncol Biol Phys. 2003;55:1209–1215
11. Paskett ED, Naughton MJ, McCoy TP, et al. The epidemiology of arm and hand swelling in premenopausal breast cancer
survivors. Cancer Epidemiol Biomarkers Prev. 2007;4:775–782
12. Goffman TE, Laronga C, Wilson L, et al. Lymphedema
of the arm and breast in irradiated breast cancer
patients: risks in an era of dramatically changing axillary surgery. Breast J. 2004;10:405–411
13. Meeske KA, Sullivan-Halley J, Smith AW, et al. Risk factors for arm lymphedema
following breast cancer
diagnosis in Black women and White women. Breast Cancer
Res Treat. 2009;113:383–391
14. Vignes S, Arrault M, Dupuy A. Factors associated with increased breast cancer
volume. Acta Oncol. 2007;46:1138–1142
15. Armer JM, Stewart BR. A comparison of four diagnostic criteria for lymphedema
in a post breast cancer
population. Lymphat Res Biol. 2005;3:208–217
16. Rönkä RH, Pamilo MS, von Smitten KA, et al. Breast lymphedema
after breast conserving treatment. Acta Oncol. 2004;43:551–557