Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors
(USC Thesis Other)
Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ARM LYMPHEDEMA IN A MUTI-ETHNIC COHORT OF FEMALE BREAST
CANCER SURVIVORS
by
Kayo Togawa, MPH
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)
May 2014
Copyright 2014 Kayo Togawa
ii
Table of Contents
List of tables iii
List of figures iv
Abstract
v
Chapter 1: Overview of breast cancer-related arm lymphedema 1
References 5
Chapter 2: Risk factors for self-reported lymphedema among female
breast cancer survivors
7
Materials and Methods 8
Results 15
Discussion 19
References 27
Chapter 3: Symptoms of breast cancer-related lymphedema and quality
of life among disease-free female breast cancer survivors
39
Materials and Methods 40
Results 47
Discussion 51
References 56
Chapter 4: The associations of lymphedema and lymphedema-
associated infection with C-reactive protein and serum amyloid A
67
Materials and Method 68
Results 72
Discussion 73
References 75
iii
List of Tables
Table 1: Participant characteristics by race/ethnicity 33
Table 2: Participant characteristics and their associations with
lymphedema among 666 female breast cancer survivors
35
Table 3: Participant characteristics and their associations with early
onset and late onset lymphedema
37
Table 4: Risk factors for arm lymphedema stratified by race/ethnicity 38
Table 5: Characteristics of 600 women by number of self-reported
lymphedema symptoms
60
Table 6: Comparison of quality of life measures by number of self-
reported lymphedema symptoms
62
Table 7: The associations between self-reported lymphedema symptoms
and physical health-related quality of life score
63
Table 8: The associations between self-reported lymphedema symptoms
and mental health-related quality of life score
64
Table 9: The associations between self-reported lymphedema symptoms
and perceived stress
65
Table 10: Self-reported lymphedema symptoms by race/ethnicity and
age at diagnosis
66
Table 11: Characteristics of 693 female breast cancer survivors by
lymphedema
78
Table 12: Geometric means of C-reactive protein (CRP) and serum
amyloid A (SAA) by lymphedema status
80
iv
List of Figures
Figure 1: Recruitment flow and data collection 31
Figure 2: Kaplan-Meier cumulative incidence estimate of self-reported
arm lymphedema
32
Figure 3: Recruitment flow and data collection 59
Figure 4: Recruitment flow and data collection 77
v
Abstract
Lymphedema remains a significant issue among breast cancer survivors and thus, further
efforts are needed to improve prevention and management of this complication. This
dissertation explores lymphedema risk factors and the associations of lymphedema with
quality of life (QOL) and inflammatory markers using data from a multi-center, multi-
ethnic prospective cohort study of breast cancer survivors (Health, Eating, Activity and
Lifestyle Study). A total of 1,183 women diagnosed with in situ or Stage I-IIIA breast
cancer between 1995 and 1999 were recruited through the Surveillance, Epidemiology,
and End Results registries in New Mexico, Los Angeles County, and Western
Washington. Data were collected on sociodemographic factors, clinical characteristics,
comorbidities, body mass index (BMI), hormonal factors, lifestyle factors, and QOL. A
blood sample was collected to assess the levels of C-reactive protein (CRP) and serum
amyloid A (SAA) at about 30 months after diagnosis. The results indicated that extensive
surgery/axillary lymph node dissection, obesity, chemotherapy, and hypertension
increased the risk of lymphedema, and that lymphedema-associated burning pain and
tension decreased QOL. No significant associations between lymphedema and CRP/SAA
were found after adjusting for BMI. These findings suggest the importance of monitoring
survivors with lymphedema risk factors and relieving lymphedema-associated burning
pain and tension.
Keywords Lymphedema, Arm morbidity, Risk factors, Breast cancer survivors, Breast
cancer treatment, Axillary lymph node dissection, Chemotherapy, Hypertension, Body
mass index, quality of life, inflammatory marker, C-reactive protein, serum amyloid A
1
Chapter 1
Overview of breast cancer-related arm lymphedema
More than 2.9 million women with a history of breast cancer were living in US in
January 2012 (1). Given that overall breast cancer incidence rates have remained
relatively stable since 2004 and breast cancer mortality has been declining steadily (1),
the number of breast cancer survivors is expected to continually increase. As more
women live longer after breast cancer, studies of breast cancer survivorship have become
greatly important.
Breast cancer survivors continue to experience various challenges after cancer and its
active treatment. Schmitz and colleagues (2) found that more than 60% of women
experienced at least one complication within six years of their breast cancer diagnoses.
One of the potentially debilitating complications experienced by breast cancer survivors
is lymphedema. This chapter summarizes the current knowledge about this complication.
Definition of lymphedema
The International Society for Lymphology has defined lymphedema as “an abnormal
collection of excessive tissue proteins, edema, chronic inflammation, and fibrosis” (3).
Lymphedema is classified into primary or secondary. Primary lymphedema can be
sporadic, hereditary, or syndrome-associated whereas secondary lymphedema is acquired
(1,4). The majority of lymphedema cases seen in developed countries are secondary
2
following cancer treatment (1,5).
Physiology of lymphedema
The lymphatic system plays a critical role in the maintenance of tissue fluid homeostasis
(6). Normally, about 90% of fluid that cross arterial wall of capillaries and enters the
interstitial space re-enters the circulatory system through the venous wall of capillaries,
and about 10% of fluid enter the lymphatic system (2,7) in order to maintain tissue fluid
homeostasis. However, when the lymphatic system is damaged, its ability to transport
lymphatic fluid is compromised causing an excess of fluid in the interstitial spaces. In
breast cancer patients, treatments such as axillary dissection, radiotherapy, or a
combination of these can damage the lymphatic system (3,7-9), resulting in an
accumulation of protein-rich interstitial fluid in the affected limb.
Prevalence and incidence of lymphedema among breast cancer survivors
In studies of breast cancer survivors, the reported prevalence or incidence of lymphedema
varies widely ranging from 2% to 56% (10). This wide range is likely due to the high
variations across studies in definition and measures of lymphedema, length of follow-up,
and associated clinical characteristics. For instance, some studies have relied on self-
report whereas others have used objective measures such as arm circumference and arm
volume. Furthermore, lack of information on time to onset of lymphedema may have
decreased the accuracy of the incidence estimates in some of the studies. Because
multiple factors influence the estimates of prevalence and incidence, it is important to
consider study methods and populations carefully when comparing the results across
3
studies or generalizing the results.
Risk factors for breast cancer-related lymphedema
Studies have evaluated a variety of demographic, health, and clinical characteristics in
relation to lymphedema. The existing literatures agree that dissection of axillary lymph
nodes and axillary radiation therapy are associated with the development of lymphedema.
In addition, factors such as obesity, surgical wound infection, tumor stage, and extent of
surgery have been consistently found to be associated with lymphedema risk (11).
However, the impact of other factors such as age, chemotherapy, and comorbid medical
conditions on lymphedema risk have varied across studies. Thus, further research is
needed to clarify inconsistent findings. Furthermore, most studies lacked racial-ethnic
diversity. It is important to obtain data from a multi-racial/ethnic population because such
data will allow us to evaluate potential heterogeneity across different racial/ethnic groups.
Impact of breast cancer-related lymphedema
Although lymphedema is not considered fatal, there are many ways through which this
condition can negatively affect women’s survivorship. Lymphedema usually presents
with symptoms such as arm puffiness; tired, thick, or heavy feeling in the arm; tightness
of clothes; change in skin texture; and pain which contrast with the woman’s unaffected
contralateral side (12). Although treatments are available to alleviate lymphedema
symptoms, many women remain undertreated due to lack of recognition of lymphedema
by health care professionals, cost of care, or lack of access to care (4,13,14). The
consequences of lymphedema include cosmetic deformity, physical discomfort and
4
upper extremity disability (15), and decreased quality of life (QOL) (16). Lymphedema
also increases the risk of infectious complications such as cellulitis and lymphangitis
because the surplus of lymph fluid and accumulated protein encourages bacterial growth,
and the dysfunction of lymphatic system compromises the immune system (8).
Furthermore, although it is a rare event, chronic lymphedema can be complicated by a
potentially fatal malignant tumor such as lymphangiosarcoma (17). In order to avoid
these debilitating consequences, research or clinical practice should focus on reducing the
risk of lymphedema and also providing effective and timely treatment for breast cancer
survivors with this complication.
The aims of this dissertation
Lymphedema continues to be a relatively common and potentially debilitating
complication after breast cancer treatment and thus, further efforts are needed to reduce
the burden of this complication in the growing population of breast cancer survivors. This
dissertation presents data from the Health, Eating, Activity and Lifestyle study to address
three questions. First, this dissertation presents data on risk factors for overall, early, and
late onset lymphedema in a multiethnic cohort of female breast cancer survivors. Second,
data on the associations of lymphedema symptoms with various aspects of the survivor’s
quality of life are presented. Finally, data on the associations between two inflammatory
markers, C-reactive protein and serum amyloid A, and lymphedema are evaluated in a
cohort of female breast cancer survivors.
5
References
1. American Cancer Society. Breast Cancer Facts & Figures 2013-2014. 2013;1–40.
2. Schmitz KH, Speck RM, Rye SA, et al. Prevalence of breast cancer treatment
sequelae over 6 years of follow-up. Cancer. 2012;118(S8):2217–2225.
3. Brennan MJ. Lymphedema following the surgical treatment of breast cancer: a
review of pathophysiology and treatment. J Pain Symptom Manage.
1992;7(2):110–116.
4. Rockson SG, Rivera KK. Estimating the Population Burden of Lymphedema.
Annals of the New York Academy of Sciences. 2008;1131(1):147–154.
5. Rourke LLL, Hunt KKK, Cormier JNJ. Breast cancer and lymphedema: a current
overview for the healthcare provider. CORD Conference Proceedings.
2010;6(3):399–406.
6. Rockson SG. Update on the Biology and Treatment of Lymphedema. Curr Treat
Options Cardio Med. 2012;14(2):184–192.
7. Fu MR, ACNS-BC SHR. The pathophysiology of lymphedema. Nursing.
2009;48–54.
8. Rockson SG. Lymphedema. Am J Med. 2001;110(4):288–295.
9. Sakorafas GH, Tsiotou AG, Balsiger BM. Axillary lymph node dissection in breast
cancer--current status and controversies, alternative strategies and future
perspectives. Acta Oncol. 2000;39(4):455–466.
10. Morrell RM, Halyard MY, Schild SE, et al. Breast cancer-related lymphedema.
Mayo Clin. Proc. 2005;80(11):1480–1484.
11. Williams AF, Franks PJ, Moffatt CJ. Lymphoedema: estimating the size of the
problem. palliat med. 2005;19(4):300–313.
12. Norman SA, Miller LT, Erikson HB, et al. Development and validation of a
telephone questionnaire to characterize lymphedema in women treated for breast
cancer. Phys Ther. 2001;81(6):1192–1205.
13. Armer JM, Radina ME, Porock D, et al. Predicting breast cancer-related
lymphedema using self-reported symptoms. Nurs Res. 2003;52(6):370–379.
14. Shih YCT, Xu Y, Cormier JN, et al. Incidence, Treatment Costs, and
Complications of Lymphedema After Breast Cancer Among Women of Working
Age: A 2-Year Follow-Up Study. Journal of Clinical Oncology.
6
2009;27(12):2007–2014.
15. Petrek JA, Pressman PI, Smith RA. Lymphedema: current issues in research and
management. CA-ATLANTA. 2000;
16. Ahmed RL, Prizment A, Lazovich D, et al. Lymphedema and Quality of Life in
Breast Cancer Survivors: The Iowa Women's Health Study. Journal of Clinical
Oncology. 2008;26(35):5689–5696.
17. Ocaña A, Delgado C, Rodríguez CA, et al. Case 3. Upper limb
lymphangiosarcoma following breast cancer therapy. Journal of Clinical
Oncology. 2006;24(9):1477–1478.
7
Chapter 2
Risk factors for self-reported arm lymphedema among female breast cancer
survivors
Lymphedema is a relatively common and potentially debilitating condition in which there
is an excessive accumulation of lymphatic fluid in the arm or hand. It develops in
approximately 20% of women after treatment for breast cancer (1,2); it can occur as a
result of damage to the lymphatic system from breast cancer treatment such as axillary
lymph node dissection or axillary radiotherapy (3). The use of sentinel lymph node
biopsy, which avoids unnecessary axillary lymph node surgery in patients with
pathologically negative nodes, has reduced the risk of lymphedema (4); however, the risk
of lymphedema has not been completely eliminated (5). About 25% of patients who
undergo sentinel lymph node biopsy have positive nodes, for which patients undergo
axillary treatment (6). Thus, many breast cancer patients remain at risk of lymphedema.
Although lymphedema is not considered life threatening, its consequences include
cosmetic deformity, physical discomfort, and upper extremity disability (7).
Lymphedema also increases the risk of cellulitis, lymphangitis, and occasionally
lymphangiosarcoma (8-10). No known definitive cure for lymphedema is available and
thus, affected women live with lymphedema for many years. Given the negative impact
on quality of life (11-14) and the potentially higher medical cost of managing
lymphedema and treating lymphedema-induced conditions (8), preventive measures are
8
desired.
Many studies have evaluated a variety of demographic, health, and clinical characteristics
in relation to lymphedema (12,15-29); however, the results vary widely, possibly due to
differences in study design, statistical power, analytic methods, measures used to define
lymphedema, demographic characteristics, or length of follow-up. Furthermore, the
majority of prior studies lacked racial/ethnic diversity (18,19,22,28,29) or long-term
follow-up data (16-19,24,27,28,30); data from long-term prospective cohort studies in a
diverse population remain scarce. The Health, Eating, Activity, and Lifestyle (HEAL)
Study, has followed a cohort of breast cancer survivors for more than 10 years and
consists of non-Hispanic white women, Hispanic women, and black women. Here, we
assess the incidence of self-reported lymphedema, timing of lymphedema onset, and
associations with breast cancer-related and treatment-related factors, sociodemograhpic
factors, comorbidities, hormone-related factors, and lifestyle factors across three racial-
ethnic groups.
MATERIALS AND METHODS
Study setting, subjects, and recruitment
The aims, study design, and recruitment procedures of the HEAL Study have been
published previously (31,32). Briefly, the HEAL Study is a multi-center, multi-ethnic
prospective breast cancer cohort study. Women diagnosed with first primary in situ or
Stage I-IIIA invasive breast cancer between 1995 and 1999 were recruited into the HEAL
Study through the Surveillance, Epidemiology, and End Results (SEER) registries in
9
three regions of the United States: New Mexico, Western Washington, and Los Angeles
County, California. The age ranges studied varied by study site with women aged under
92 years recruited in New Mexico, women aged 40 to 65 years recruited in Western
Washington, and women aged 35 to 64 years recruited in Los Angeles County. A total of
1,183 women completed up to five assessments over 10 years of follow up. Four of those
assessments were used for this study. The first assessment (baseline assessment) was
administered in person within the first year (on average, 6 months) after a woman’s
diagnosis (Figure 1). The second assessment was administered, on average, 30 months
after a woman’s diagnosis (30-month assessment). The 30-month assessment was
administered via in-person interview or self-completed questionnaire. The third
assessment was administered, on average, 40 months after a woman’s diagnosis (40-
month assessment) by telephone interview or mailed questionnaire in New Mexico, by
mailed questionnaire plus telephone follow-up in Western Washington, and by telephone
interview in Los Angeles County. The last assessment was administered, on average, 123
months after a woman’s diagnosis (123-month assessment) by telephone interview in
New Mexico and Los Angeles County and by mailed questionnaire or telephone
interview in Western Washington.
In the present study, we excluded 217 women who were younger than 35 years (n=4) or
older than 64 years (n=213) in order to provide similar age distributions across study
sites. We also excluded 14 women who did not receive any type of surgery; and 32
women whose racial/ethnic classification was other than Hispanic white, non-Hispanic
white, or black, leaving 920 women. A total of 751 (82%) of the 920 women completed
10
the 30-month assessment. A total of 688 (92%) of the 751 women completed either the
40-month assessment or the 123-month assessment (496 women completed both the 40-
month and the 123-month assessments, 163 women completed only the 40-month
assessment, 29 women completed only the 123-month assessment). We then excluded 22
women with incomplete or questionable data on onset date of lymphedema (n=12),
number of excised lymph nodes (n=4), or body mass index (BMI) (n=6). The final
analytic cohort consisted of 666 women; 666 (72%) of the eligible cohort had complete
data on all assessments used in this analysis (Figure 1).
We obtained informed consent from all participants at each assessment. The study was
approved by the institutional review boards of participating centers, in accord with
assurances filed with and approved by the U.S. Department of Health and Human
Services.
Data collection
Arm lymphedema. Participants provided information on lymphedema at the 40-month and
123-month assessments. To determine the presence of lymphedema, we first presented
our definition of lymphedema to study participants: “Sometimes the arm on the side on
which you had breast cancer becomes swollen because of an accumulation of fluid in
your arm. This is called lymphedema. Please do not confuse this with the temporary
swelling that occurs after surgery.” We then asked the following “yes” or “no” question:
“Have you experienced lymphedema in your arm at any time since your breast cancer
diagnosis?” For women who answered “yes,” we also asked when they first experienced
11
lymphedema symptoms (month and year), and whether they were still experiencing
lymphedema at the time of assessment. Lymphedema occurring within one year of breast
cancer diagnosis was considered early onset lymphedema, whereas that occurring more
than one year after diagnosis was considered late onset lymphedema. Note that we have
previously reported on factors associated with development of lymphedema at the 40-
month assessment among the black patients, all of whom were from Los Angeles County
(2). In the aforementioned report, the black breast cancer patients in the HEAL Study
were compared to a group of non-Hispanic white breast cancer patients who were not
participants in the HEAL Study. The present study provides additional follow-up for
these black patients and includes data on the other eligible HEAL Study participants.
Breast cancer-related and treatment-related factors. We obtained clinical data on
diagnosis date, age at diagnosis, disease stage (SEER staging), treatment types (surgeries,
radiation therapy, chemotherapy), tumor size, cancer location, and number of excised
lymph nodes from SEER cancer registry records and by abstracting participants’ hospital
medical records. Tamoxifen use at or prior to baseline was assessed using both hospital
medical records and self-report.
Sociodemographic and lifestyle factors. The baseline assessment captured
sociodemographic information such as race/ethnicity, marital status, educational status,
and insurance status. The baseline assessment also captured information on personal
histories of smoking, alcohol intake in the year prior to diagnosis, and sports and
recreational physical activity in the year prior to diagnosis. We calculated pack-years of
smoking as the number of packs of cigarettes per day times the number of years the
12
woman smoked, grams of alcohol consumed per day, and metabolic equivalent task
(MET) hours of sports and recreational physical activity based on the Compendium of
Physical Activities compiled by Ainsworth et al (33).
Health-related and hormone-related factors. Information on comorbid medical
conditions such as diabetes, hypertension, and arthritis, was collected through self-report
and hospital records. Charlson Comorbidity Index (34) was calculated based on data from
hospital medical records (35). We did not count any carcinoma as part of the Charlson
Comorbidity Index. Participants also reported information on height at age 18 years and
weight five years before diagnosis (black women in Los Angeles County) or weight one
year before diagnosis (Hispanic women in New Mexico and Non-Hispanic white women
in New Mexico and Western Washington) at the baseline assessment. BMI prior to
diagnosis was calculated as weight (kg) divided by the square of height (m
2
). The
baseline assessment also captured information on oral contraceptive use and
postmenopausal hormone replacement therapy use before breast cancer diagnosis, and
menopausal status at diagnosis. Menopausal status was determined based on the
following questionnaire data: age, date of last menstruation, and hysterectomy and
oophorectomy status (36).
Statistical analysis
Descriptive statistics by race/ethnicity were obtained for breast cancer-related and
treatment-related factors, sociodemographic factors, health-related factors, hormonal
factors, and lifestyle factors (Table 1). Time to onset of lymphedema was calculated as
13
the time from breast cancer diagnosis until self-reported first lymphedema occurrence;
we used these data to generate a cumulative incidence curve (Figure 2).
To identify factors associated with self-reported lymphedema, we fit Cox proportional
hazards models and obtained estimates of the hazard ratio (HR) and its 95% confidence
interval (CI) using time since diagnosis as the time scale (37). Women were followed
from date of breast cancer diagnosis to date of first lymphedema occurrence or until date
of last follow up. One hundred fifty-six women who died or were lost to follow-up
between the 40-month and the 123-month assessments contributed follow up only
through the 40-month assessment because their lymphedema status after the 40-month
assessment was unknown. We considered the following variables as potential covariates
in our multivariate models: breast cancer disease stage (in situ, localized, regional), tumor
size (<10, 10 to 19, 20 mm or greater, missing), tumor location (nipple/central portion or
upper/lower inner quadrant, upper/lower outer quadrant, overlapping lesion, axillary tail
or not specified), surgery type (partial or less than total mastectomy or unspecified
surgery, total or modified radical mastectomy), reconstructive surgery (yes, no), number
of excised lymph nodes (0, 1 to 9, 10 or more), radiation therapy (yes, no), chemotherapy
(yes, no), tamoxifen (yes, no), marital status (married, widowed, divorced or separated,
never married), education status (high school or less, some college, college degree,
graduate studies), medical insurance (yes, no), BMI prior to diagnosis [<25 (underweight
or normal), 25 to 29.9 (overweight), and 30 kg/m
2
or above (obese)], menopausal status
at diagnosis (premenopausal, postmenopausal, unknown), hypertension at or prior to
diagnosis (yes, no), diabetes at or prior to diagnosis (yes, no), arthritis at or prior to
14
diagnosis (yes, no), and Charlson Comorbidity Index (0, 1 to 2), oral contraceptive use
prior to diagnosis (yes, no), estrogen use prior to diagnosis (yes, no, missing), progestin
use prior to diagnosis (yes, no, missing), pack-years of smoking (less than 0.05, 0.05 to
5.3, 5.4 to 20.5, more than 20.5), alcohol intake in the year prior to diagnosis (less than 1,
1 to 6, more than 6 grams per day), physical activity during the year prior to diagnosis
(less than 0.5, 0.5 to 20.0, more than 20 MET hours per week). Age at diagnosis (35 to
44, 45 to 49, 50 to 54, 55 to 59, 60 to 64 years) and race/ethnicity [(non-Hispanic white
(Western Washington, New Mexico), black (Los Angeles County), Hispanic white (New
Mexico)] were considered design variables and were included in all models.
Each variable was added to an age- and race/ethnicity- adjusted model and a likelihood
ratio test was performed to test whether the variable significantly improved the model fit.
Postmenopausal estrogen use and progestin use were examined among postmenopausal
women only. In analyses of postmenopausal women and menopausal status, age was
treated as continuous. All variables with a p value less than 0.05 based on the likelihood
ratio test were then added to an age- and race/ethnicity- adjusted model one at a time in
order from the variable with the smallest p value to the variable with largest p value to
examine whether the addition of each variable improved the fit of the model. Only the
variables that significantly improved the model were kept in the final model. The final
multivariable model included race/ethnicity, age at diagnosis, surgery type, number of
excised lymph nodes, chemotherapy, BMI and hypertension. We also tested for exposure-
associated trends in risk of lymphedema for variables represented by ordinal values by
fitting the original values in our model and testing whether the coefficient associated with
15
that covariate differed from zero. In addition, we tested interactions between variables
included in the final multivariable model by creating an interaction term and using a
likelihood ratio test. We considered a two-sided p value less than 0.05 as statistically
significant.
In exploratory analyses, we evaluated whether the association with lymphedema for each
of the variables in the final multivariable model differed by timing of lymphedema onset;
to do this, we created an interaction term using a time-dependent indicator for follow-up
(12 months or less, longer than 12 months). We then performed a likelihood ratio test to
compare the interaction model to the model without the interaction term. We also
assessed the association of each risk factor with the development of lymphedema within
each racial/ethnic group. Finally, we performed sensitivity analyses by excluding women
with in situ breast cancer and reevaluating the final multivariable model; we also reran
analyses censoring women who developed a recurrence or new primary breast cancer at
the time of these events. The results were not meaningfully changed.
All analyses were performed using the STATA software (Version 12; College Station,
TX, USA).
RESULTS
Characteristics of study population. The study population consisted of three racial/ethnic
groups: 371 non-Hispanic white women [n=225 (New Mexico), n=146 (Western
Washington)], 226 black women (Los Angeles County), and 69 Hispanic women (New
Mexico). Participant characteristics by race/ethnicity are shown in Table 1. The mean
16
age at breast cancer diagnosis was 51.5 years (standard deviation=7.2). Approximately
23% of the participants had in situ breast cancer, 52% had localized disease, and 25% had
regional disease. Prior to breast cancer diagnosis, approximately 33% of the women were
overweight (BMI=25.0-29.9 kg/m
2
) and 21% were obese (BMI>30 kg/m
2
).
The median length of follow-up since breast cancer diagnosis was 10.2 years (range: 2.2-
12.4). During follow-up, 190 women (29%) reported lymphedema. The median time
from breast cancer diagnosis to onset of lymphedema was 10.5 months (range: 0.5-134.9
months). The cumulative incidence of lymphedema was 15.8%, 20.9%, 26.1%, and
29.7%, at 1, 2, 5, and 10 year(s), respectively (Figure 2). One hundred nine (77%) of 141
women who reported lymphedema at the 40-month assessment completed the 123-month
assessment. Among these 109 women, 63 (58%) women indicated that their lymphedema
was still present at the time of the 123-month assessment.
Age- and race/ethnicity-adjusted analyses. We present a series of hazard ratios for
individual factors considered as potential risk factors for lymphedema in Table 2. When
age at diagnosis and race/ethnicity were mutually adjusted, the oldest age group (60 to 64
years) had a lower risk of lymphedema than the youngest age group (35 to 44 years)
(HR=0.59, 95% CI: 0.35, 0.97) and black women had a higher risk of lymphedema than
non-Hispanic white women (HR=1.62, 95% CI: 1.20, 2.20). Women with in situ breast
cancer had a substantially lower risk of lymphedema than women with localized breast
cancer after adjusting for age at diagnosis and race/ethnicity (HR=0.26, 95% CI: 0.15,
0.45). Women with regional breast cancer were more likely to develop lymphedema than
women with localized breast cancer; however the CI included 1.0 (HR=1.35, 95% CI:
17
0.99, 1.85). We excluded disease stage from the final model since disease stage was
strongly associated with treatment-related factors. In addition to disease stage, the
following variables added significantly to the age- and race/ethnicity-adjusted model
based on likelihood ratio test (P<0.05): tumor size, surgery type, number of excised
lymph nodes, chemotherapy, tamoxifen, BMI, and hypertension. No statistically
significant association was observed for other factors.
Multivariable analyses overall. Tumor size and tamoxifen did not significantly improve
the model fit once number of excised lymph nodes was in the model; hence these
variables were not included in the final multivariable model. The final model included
age at diagnosis, race/ethnicity, surgery type, number of excised lymph nodes,
chemotherapy, BMI, and hypertension. In this multivariable model, the 95% CI for black
women no longer excluded one (vs. non-Hispanic white women; HR=1.16, 95% CI: 0.83,
1.61). Treatment-related factors including total or modified radical mastectomy (vs.
partial or less than total mastectomy; HR=1.37, 95% CI: 1.01, 1.85), 10 or more excised
lymph nodes (vs. none removed; HR=6.05, 95% CI: 3.01, 12.16), and chemotherapy
(HR=1.48, 95% CI: 1.09, 2.02) increased the risk of lymphedema. The risk of developing
lymphedema increased by 5% for each lymph node removed (P-trend<0.001). Among the
health-related factors, having BMI ≥30 kg/m
2
(vs. BMI<25 kg/m
2
; HR=1.59, 95% CI:
1.09, 2.31) and having hypertension (HR=1.49, 95% CI: 1.06, 2.10) were associated with
increased lymphedema risk. We observed a statistically significant increasing trend in
risk with increasing BMI (P-trend=0.01). None of the pre-diagnosis lifestyle factors had a
substantial effect on the risk of lymphedema. We found statistically significant
18
interactions between race/ethnicity and chemotherapy (P=0.01) and hypertension and
chemotherapy (P=0.004) among the 511 women with invasive breast cancer, and
between race/ethnicity and hypertension (P=0.001) among all women (data not shown).
Multivariable analyses by timing of lymphedema onset. Women’s characteristics and
their associations with early onset and late onset lymphedema are presented in Table 3.
Older age was associated with decreased risk of late onset lymphedema (P-trend=0.002),
but not with early onset lymphedema (P-trend=0.91). The risk of late onset lymphedema
decreased by 5% per year increase in age (HR=0.95, 95% CI: 0.92, 0.98). We observed a
statistically significant difference in the age effect between early and late onset
lymphedema when age was treated as continuous (P=0.02). Our results also showed that
the risk of both early onset and late onset lymphedema increased with increasing number
of excised lymph nodes (HR=1.04, 95% CI: 1.01, 1.06, HR=1.06, 95% CI: 1.03, 1.09,
respectively). The influence of chemotherapy on risk for late onset lymphedema
(HR=1.77, 95% CI: 1.13, 2.78) was greater than the effect of chemotherapy on risk for
early onset lymphedema (HR=1.28, 95% CI: 0.86, 1.91); however, these two hazard
ratios did not differ statistically (P=0.85).
Multivariable analyses stratified by race/ethnicity. The results showed that hypertension
was associated with an elevated risk of lymphedema only among black women
(HR=2.73, 95% CI: 1.65, 4.53) (Table 4). Similarly, receipt of chemotherapy increased
the risk of lymphedema among black women (HR=2.69, 95% CI: 1.61, 4.50), but not
other women.
19
DISCUSSION
This prospective cohort study of women diagnosed with first primary in situ or Stage I-
IIIA invasive breast cancer between the ages of 35 and 64 years supports previous
findings that the risk of lymphedema is higher among women who had more lymph nodes
removed, more extensive surgery, and higher BMI. This study also highlights the
importance of long-term monitoring of breast cancer survivors, particularly those who are
younger, have had more lymph nodes removed, or received chemotherapy, as they are at
a higher risk of developing late onset lymphedema.
This study demonstrates that the cumulative incidence of lymphedema increases with
time and that lymphedema can develop later in the survival trajectory. Although many
studies have reported incidence of lymphedema, comparison across studies is difficult
because of variability in lymphedema definition and assessment, length of follow-up, and
associated patient characteristics. In a study conducted by Kwan and colleagues (17)
where length of follow-up for each participant was considered to calculate cumulative
incidence of lymphedema, the cumulative incidence of lymphedema was 10.4% at one
year and 13.5% at two years after breast cancer diagnosis. These cumulative incidence
values were lower than those observed in our study, perhaps because their participants
were diagnosed with breast cancer more recently when sentinel lymph node biopsy was
more common than it was when participants in HEAL Study were diagnosed. Also, the
study by Kwan et al required hospital records to establish the presence of lymphedema.
The existing data on whether lymphedema incidence varies by age have been
20
inconsistent. Contrary to some of the previous studies (16,18) and despite the limited age
range in our study (35 to 64 years), we found that older women were less likely to
develop lymphedema than younger women. We observed a statistically significant,
decreasing linear trend in lymphedema risk associated with increasing age after adjusting
for breast cancer characteristics and treatment factors, hypertension, and BMI (P=0.04).
Our results for age are consistent with those from other studies (19,29,38). When we
examined the effects of age by timing of onset, we found that age was inversely
associated with risk of late onset lymphedema, but not with risk of early onset
lymphedema. The age-by-time interaction was statistically significant when age was
treated as continuous (P=0.02). The reason why older women are less likely to develop
lymphedema later in the survival trajectory is unclear. A previous study showed that
infection or injury was associated with increased risk of late onset lymphedema (20). We
were unable to adjust for the effect of infection or injury, thus we cannot rule out the
possibility of residual confounding. More research is needed to study the age effect on
late onset lymphedema.
This study identified two modifiable risk factors for lymphedema. One of the modifiable
risk factors was pre-diagnosis BMI; women with higher pre-diagnosis BMI were at a
higher risk of lymphedema. The association between pre-diagnosis BMI and risk of
lymphedema has been shown in the Iowa Women’s Health Study (22) and another study
conducted by Jammallo et al (30). Other studies have found an association between at-
diagnosis or post-diagnosis BMI and the risk of lymphedema (15,23,24,28). The
mechanism by which excess weight increases the risk of lymphedema remains unclear;
21
the risk of lymphedema may be elevated in obese women due to additional demand on
both the vascular and lymphatic systems to transport fluid (28). Weight gain during
survivorship in relation to lymphedema risk is also of interest. Petrek et al (20) reported
an association between weight gain since operation and risk of breast cancer-related
lymphedema. However, our study failed to demonstrate an association between change in
BMI (BMI at 30-month assessment minus pre-diagnosis BMI) and late onset
lymphedema (data not shown). More studies are needed to understand the association
between BMI and the risk of lymphedema.
The other modifiable risk factor we found in the present study was hypertension. We
found that hypertension was associated with increased risk of lymphedema after adjusting
for BMI, which contradicts some studies where no association between hypertension and
lymphedema was found (21-23,27). A possible mechanism by which hypertension
increases the risk is through increased capillary filtration due to elevated hydrostatic
pressure. When capillary filtration increases and lymphatic drainage is insufficient, the
fluid accumulates in interstitial space leading to swelling of the arm. We found no
significant difference in lymphedema risk between women with hypertension who were
taking antihypertensive medications and those who were not (P=0.74). When we
examined different types of antihypertensive medications, we found no increase in the
risk of lymphedema among women with hypertension who were taking at least one of the
following antihypertensive medications: vasodilators, calcium-blockers, beta-blockers,
alpha-blockers, or angiotensin-converting-enzyme inhibitors at the time of 30-month
assessment (vs. women without hypertension; HR=1.21, 95% CI: 0.70, 2.07) while the
22
risk was elevated among women with hypertension who were taking diuretics (vs.
women without hypertension; HR=1.86, 95% CI:1.18, 2.94). However, this elevated risk
among women who were taking diuretics may reflect the use of diuretics as treatment for
lymphedema. Due to lack of information on duration and timing of medication use in this
study, this finding will need to be clarified in future studies.
Our study results agree with the majority of studies showing that lymphedema risk
increases with increasing number of excised lymph nodes (12,17,22,25). This positive
association between number of excised lymph nodes and risk remained significant after
restricting the cohort to women with invasive breast cancer (P-trend=0.002). On the other
hand, our study failed to confirm an association with radiation therapy. We also did not
observe effect modification of radiation therapy by number of excised lymph nodes
(P=0.18). However, detailed information on location, dose, and duration of radiation
therapy was not available in this study. Since the risk of lymphedema depends on the type
of radiotherapy (39), we may have missed an association with radiation treatment.
Evidence on associations between chemotherapy and lymphedema risk has been
inconsistent, with some studies finding a positive association (8,12,17,19,22,23) and
others finding no association (25,27,29). The present study showed a positive association
between chemotherapy and overall lymphedema risk after adjusting for other risk factors
(HR=1.48, 95% CI: 1.09, 2.02). This association persisted after excluding women with in
situ breast cancer (HR=1.47, 95% CI: 1.06, 2.03). When we examined individual types of
chemotherapy adjusting for other risk factors, we found that 5-fluorouracil, methotrexate,
and cyclophosphamide were associated with increased risk of lymphedema whereas
23
taxane and doxorubicin were not. Women who received 5-fluorouracil, methotrexate, and
cyclophosphamide in combination may have a higher risk of infection because these
agents together tend to reduce the number of white blood cells and compromise immune
response (40). Therefore, it is conceivable that the elevated lymphedema risk observed in
women who received these chemotherapy agents may be partly attributable to infection.
Our exploratory analysis further demonstrated a statistically significant interaction
between chemotherapy and hypertension in relation to lymphedema risk among women
diagnosed with invasive breast cancer. More specifically, the risk of lymphedema was
elevated only among women with hypertension who received chemotherapy (vs. women
without hypertension who did not receive chemotherapy; HR=2.36, 95% CI: 1.54, 3.64).
Swystun and colleagues (41) suggested that the chemotherapy metabolite acrolein may
upregulate procoagulant pathways, while impairing endogenous anticoagulant pathways.
In this study, 222 (88%) of 253 women treated with chemotherapy received
cyclophosphamide, which has been associated with an increased risk of thrombosis (41-
43). It is possible that this procoagulant activity might overburden the lymphatic system
in women with preexisting hypertension.
Although lymphedema among breast cancer survivors has been studied extensively, data
from studies with long-term follow-up of ethnically diverse populations remain scarce.
Kwan et al (17) found that black women had a higher risk of lymphedema compared to
white women whereas Meeske et al (2), found no association between race (black vs.
non-Hispanic white) and lymphedema risk after adjusting for other risk factors. The black
women in the study by Meeske et al are the same black women who are included in this
24
report, but follow-up for lymphedema was limited to the 40-month assessment in their
study. With follow-up extended through the 123-month assessment, we observed that the
elevated lymphedema risk among black women was attenuated after adjusting for age,
treatment-related factors, BMI, and hypertension (vs. non-Hispanic white women;
HR=1.16, 95% CI: 0.83, 1.61). Our exploratory analysis examining racial-ethnic specific
associations showed that chemotherapy and hypertension were lymphedema risk factors
only among black breast cancer survivors; thus, the observed overall effects of these
factors on lymphedema risk appear to be limited to black women.
In our study, the median length of follow-up since breast cancer diagnosis was 10.2
years; this is longer than the median length of follow-up in many prior breast-cancer-
related lymphedema studies. With this extended follow-up time, we were able to capture
a substantial number of cases of late onset lymphedema which allowed us to compare risk
factors between early and late onset lymphedema. Moreover, unlike many previous
studies, this study consisted of a large, multiethnic sample of breast cancer survivors
drawn from population-based cancer registries. This diverse population allowed us to
study associations within groups defined by race and ethnicity and identify potential
heterogeneity across race and ethnicity. However, this study has several limitations which
should be considered when interpreting the results. This study consisted of women who
were diagnosed between 1995 and 1999 when sentinel lymph node biopsy was less
common. According to Chen et al (44), the use of sentinel lymph node biopsy increased
from 26.8% in 1998 to 65.5% in 2005 among early stage breast cancer patients. Given
that the risk of lymphedema is smaller when sentinel lymph node biopsy is used instead
25
of axillary lymph node dissection, our estimate of lymphedema incidence may be greater
than the incidence observed in more recent years. Furthermore, the selection of
chemotherapeutic agents has expanded over the years, possibly limiting the application of
results to more recent regimens. This study relied solely on self-report to define existence
of lymphedema and the self-report was not verified using medical records. However, both
sensitivity and specificity of self-reported lymphedema were found to be high in a
previous study (45), and our results are largely consistent with those from studies where
objective measurements were used (46), with the exception of the effects of radiotherapy
and chemotherapy. We were also limited by insufficient statistical power to study three-
way interactions; for example, we were unable to determine how race/ethnicity,
chemotherapy, and hypertension interact in relation to lymphedema risk.
26
In conclusion, our multi-ethnic cohort study confirms that lymphedema incidence is
highest in the first year following breast cancer diagnosis, but also indicates that
approximately 45% of women who developed lymphedema first experienced the
condition more than one year after initial diagnosis. Furthermore, it shows that the
majority of lymphedema cases persist for a long time. It is important to pay closer
attention to women who had extensive lymph node dissection, had more extensive
surgery, or were obese prior to diagnosis in order to detect lymphedema and provide
education and treatment to manage lymphedema at an early stage. More data from
multiethnic cohort studies are needed to confirm our finding that hypertension and
chemotherapy are risk factors only among black women. Clinical trials are needed to
determine whether treatment for hypertension and obesity might prevent the incidence or
severity of lymphedema in breast cancer survivors.
27
References
1. DiSipio T, Rye S, Newman B, et al. Incidence of unilateral arm lymphoedema
after breast cancer: a systematic review and meta-analysis. The Lancet Oncology.
2013;14(6):500–515.
2. 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. 2008;113(2):383–391.
3. Norrmen C, Tammela T, Petrova TV, et al. Biological Basis of Therapeutic
Lymphangiogenesis. Circulation. 2011;123(12):1335–1351.
4. Mansel RE, Fallowfield L, Kissin M, et al. Randomized Multicenter Trial of
Sentinel Node Biopsy Versus Standard Axillary Treatment in Operable Breast
Cancer: The ALMANAC Trial. JNCI Journal of the National Cancer Institute.
2006;98(9):599–609.
5. Nielsen I, Gordon S, Selby A. Breast cancer-related lymphoedema risk reduction
advice: A challenge for health professionals. Cancer Treatment Reviews.
2008;34(7):621–628.
6. Goyal A, Duley L, Fakis A. Axillary treatment for patients with early breast cancer
and lymph node metastasis: systematic review protocol. World J Surg Oncol.
2013;11:6.
7. Petrek JA, Pressman PI, Smith RA. Lymphedema: current issues in research and
management. CA-ATLANTA. 2000;
8. Shih YCT, Xu Y, Cormier JN, et al. Incidence, Treatment Costs, and
Complications of Lymphedema After Breast Cancer Among Women of Working
Age: A 2-Year Follow-Up Study. Journal of Clinical Oncology.
2009;27(12):2007–2014.
9. Torres Lacomba M, Yuste Sanchez MJ, Zapico Goni A, et al. Effectiveness of
early physiotherapy to prevent lymphoedema after surgery for breast cancer:
randomised, single blinded, clinical trial. BMJ. 2010;340(jan12 1):b5396–b5396.
10. Ocaña A, Delgado C, Rodríguez CA, et al. Case 3. Upper limb
lymphangiosarcoma following breast cancer therapy. Journal of Clinical
Oncology. 2006;24(9):1477–1478.
11. Ahmed RL, Prizment A, Lazovich D, et al. Lymphedema and Quality of Life in
Breast Cancer Survivors: The Iowa Women's Health Study. Journal of Clinical
Oncology. 2008;26(35):5689–5696.
12. Paskett ED, Naughton MJ, McCoy TP, et al. The Epidemiology of Arm and
28
Hand Swelling in Premenopausal Breast Cancer Survivors. Cancer Epidemiology
Biomarkers & Prevention. 2007;16(4):775–782.
13. Kwan W. Chronic Arm Morbidity After Curative Breast Cancer Treatment:
Prevalence and Impact on Quality of Life. Journal of Clinical Oncology.
2002;20(20):4242–4248.
14. Beaulac SM, McNair LA, Scott TE, et al. Lymphedema and quality of life in
survivors of early-stage breast cancer. Arch Surg. 2002;137(11):1253–1257.
15. Dominick SA, Madlensky L, Natarajan L, et al. Risk factors associated with breast
cancer-related lymphedema in the WHEL Study. J Cancer Surviv. 2012;
16. Hayes SC, Janda M, Cornish B, et al. Lymphedema after breast cancer: incidence,
risk factors, and effect on upper body function. Journal of Clinical Oncology.
2008;26(21):3536–3542.
17. Kwan ML, Darbinian J, Schmitz KH, et al. Risk factors for lymphedema in a
prospective breast cancer survivorship study: the Pathways Study. Arch Surg.
2010;145(11):1055–1063.
18. Mak SS, Yeo W, Lee YM, et al. Predictors of lymphedema in patients with breast
cancer undergoing axillary lymph node dissection in Hong Kong. Nurs Res.
2008;57(6):416–425.
19. Geller BM, Vacek PM, O'Brien P, et al. Factors associated with arm swelling after
breast cancer surgery. J Womens Health (Larchmt). 2003;12(9):921–930.
20. Petrek JA, Senie RT, Peters M, et al. Lymphedema in a cohort of breast carcinoma
survivors 20 years after diagnosis. Cancer. 2001;92(6):1368–1377.
21. Soran A, D'Angelo G, Begovic M, et al. Breast cancer-related lymphedema--what
are the significant predictors and how they affect the severity of lymphedema? The
Breast Journal. 2006;12(6):536–543.
22. Ahmed RL, Schmitz KH, Prizment AE, et al. Risk factors for lymphedema in
breast cancer survivors, the Iowa Women’s Health Study. Breast Cancer Res
Treat. 2011;130(3):981–991.
23. Norman SA, Localio AR, Kallan MJ, et al. Risk factors for lymphedema after
breast cancer treatment. Cancer Epidemiology Biomarkers & Prevention.
2010;19(11):2734–2746.
24. Helyer LK, Varnic M, Le LW, et al. Obesity is a risk factor for developing
postoperative lymphedema in breast cancer patients. The Breast Journal.
2010;16(1):48–54.
25. Herd-Smith AA, Russo AA, Muraca MGM, et al. Prognostic factors for
29
lymphedema after primary treatment of breast carcinoma. Cancer.
2001;92(7):1783–1787.
26. Powell SN, Taghian AG, Kachnic LA, et al. Risk of lymphedema after regional
nodal irradiation with breast conservation therapy. International Journal of
Radiation Oncology*Biology*Physics. 2003;55(5):1209–1215.
27. Ozaslan C, Kuru B. Lymphedema after treatment of breast cancer. Am. J. Surg.
2004;187(1):69–72.
28. Ridner SH, Dietrich MS, Stewart BR, et al. Body mass index and breast cancer
treatment-related lymphedema. Support Care Cancer. 2011;19(6):853–857.
29. Clough-Gorr KM, Ganz PA, Silliman RA. Older Breast Cancer Survivors: Factors
Associated with Self-reported Symptoms of Persistent Lymphedema Over 7 years
of Follow-up. The Breast Journal. 2010;16(2):147–155.
30. Jammallo LS, Miller CL, Singer M, et al. Impact of body mass index and weight
fluctuation on lymphedema risk in patients treated for breast cancer. Breast Cancer
Res Treat. 2013;142(1):59–67.
31. McTiernan A. Adiposity and Sex Hormones in Postmenopausal Breast Cancer
Survivors. Journal of Clinical Oncology. 2003;21(10):1961–1966.
32. Irwin ML, Crumley D, McTiernan A, et al. Physical activity levels before and after
a diagnosis of breast carcinoma. Cancer. 2003;97(7):1746–1757.
33. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities:
an update of activity codes and MET intensities. Medicine & Science in Sports &
Exercise. 2000;32(9 Suppl):S498–504.
34. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic
comorbidity in longitudinal studies: development and validation. J Chronic Dis.
1987;40(5):373–383.
35. Harlan LC, Klabunde CN, Ambs AH, et al. Comorbidities, therapy, and newly
diagnosed conditions for women with early stage breast cancer. J Cancer Surviv.
2009;3(2):89–98.
36. Alfano CM, McGregor BA, Kuniyuki A, et al. Psychometric properties of a tool
for measuring hormone-related symptoms in breast cancer survivors. Psycho-
Oncology. 2006;15(11):985–1000.
37. Enderlein G. Cox, D. R.; Oakes, D.: Analysis of Survival Data. Chapman and Hall,
London – New York 1984, 201 S., £ 12,–. Biom. J. 1987;29(1):114–114.
38. Gärtner R, Jensen M-B, Kronborg L, et al. Self-reported arm-lymphedema and
functional impairment after breast cancer treatment--a nationwide study of
30
prevalence and associated factors. Breast. 2010;19(6):506–515.
39. Hayes SB, Freedman GM, Li T, et al. Does axillary boost increase lymphedema
compared with supraclavicular radiation alone after breast conservation? Int. J.
Radiat. Oncol. Biol. Phys. 2008;72(5):1449–1455.
40. De Maio E, Gravina A, Pacilio C, et al. Compliance and toxicity of adjuvant CMF
in elderly breast cancer patients: a single-center experience. BMC Cancer.
2005;5:30.
41. Swystun LL, Mukherjee S, Levine M, et al. The chemotherapy metabolite acrolein
upregulates thrombin generation and impairs the protein C anticoagulant pathway
in animal-based and cell-based models. J Thromb Haemost [electronic article].
2011;9(4):767–775.
(http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=21320
281&retmode=ref&cmd=prlinks)
42. Tempelhoff von GF, Dietrich M, Hommel G, et al. Blood coagulation during
adjuvant epirubicin/cyclophosphamide chemotherapy in patients with primary
operable breast cancer. J. Clin. Oncol. 1996;14(9):2560–2568.
43. Clahsen PCP, van de Velde CJC, Julien JPJ, et al. Thromboembolic complications
after perioperative chemotherapy in women with early breast cancer: a European
Organization for Research and Treatment of Cancer Breast Cancer Cooperative
Group study. J. Clin. Oncol. 1994;12(6):1266–1271.
44. Chen AY, Halpern MT, Schrag NM, et al. Disparities and Trends in Sentinel
Lymph Node Biopsy Among Early-Stage Breast Cancer Patients (1998-2005).
JNCI Journal of the National Cancer Institute. 2008;100(7):462–474.
45. Norman SA, Localio AR, Potashnik SL, et al. Lymphedema in breast cancer
survivors: incidence, degree, time course, treatment, and symptoms. Journal of
Clinical Oncology. 2009;27(3):390–397.
46. Tsai RJ, Dennis LK, Lynch CF, et al. The Risk of Developing Arm Lymphedema
Among Breast Cancer Survivors: A Meta-Analysis of Treatment Factors. Ann Surg
Oncol. 2009;16(7):1959–1972.
31
Figure 1. Recruitment flow and data collection
Data collection: Timing and variables
Recruitment flow
Baseline assessment
(on average, 6 months following breast cancer
diagnosis)
- Sociodemographics: race/ethnicity, marital
status, education, insurance
- BMI, menopausal status
- Hormonal factors: oral contraceptive use,
hormone therapy use
- Lifestyle factors: smoking, alcohol intake,
physical activity
Hospital medical records (baseline data)/SEER
cancer registry records
- Age at breast cancer diagnosis
- Breast cancer stage, tumor size and location
- Treatment (surgery, number of excised lymph
nodes, radiation, chemotherapy, tamoxifen)
- Comorbidity
1,183 participants completed the baseline
assessment.
- 615 New Mexico
- 202 Western Washington
- 366 Los Angeles County
751 of the 920 participants who completed
the baseline assessment completed the
second assessment.
Exclusions, n (reason):
- 33 (deceased)
- 136 (did not complete the second
assessment)
Second assessment
(on average, 30 months following breast cancer
diagnosis)
- Comorbidity
- Medication use
Third assessment (on average, 40 months
following breast cancer diagnosis)/Fourth
assessment (on average, 123 months following
breast cancer diagnosis)
- Self-reported presence or absence of
lymphedema
- Self-reported onset date of lymphedema
666 of the 688 participants who completed
the third or fourth assessment comprised
the analytic cohort.
Exclusions, n (reason):
- 12 (no date of lymphedema onset)
- 4 (suspicious information on
number of excised lymph
nodes)
- 6 (incomplete information on
weight or height)
688 of the 751 participants who completed
the second assessment completed the third
or fourth assessment.
Exclusions, n (reason):
- 10 (deceased)
- 53 (did not complete the third or
fourth assessment)
920 of the 1,183 participants who completed
the baseline assessment met the criteria for
the analysis.
Exclusions, n (reason):
- 4 (age <35 years)
- 213 (age ≥65 years)
- 14 (no surgery)
- 32 (race is other than
non-Hispanic white, black, and
Hispanic white)
32
Figure 2. Kaplan-Meier cumulative incidence estimate of self-reported arm lymphedema
33
Table 1. Participant characteristics by race/ethnicity
Characteristics Categories
Non-Hispanic white
(N=371)
%
Black
(N=226)
%
Hispanic white
(N=69)
%
Breast cancer and treatment
Age at diagnosis (years) 35–44 56 (15.1) 53 (23.5) 14 (20.3)
45–49 65 (17.5) 46 (20.4) 17 (24.6)
50–54 112 (30.2) 48 (21.2) 15 (21.7)
55–59 77 (20.8) 37 (16.4) 14 (20.3)
60–64 61 (16.4) 42 (18.6) 9 (13.0)
Mean (standard deviation) 52.2 (6.7) 50.8 (7.8) 50.8 (7.4)
Disease stage In situ 96 (25.9) 47 (20.8) 12 (17.4)
Localized 206 (55.5) 98 (43.4) 42 (60.9)
Regional 69 (18.6) 81 (35.8) 15 (21.7)
Tumor size (mm) <10 113 (30.5) 45 (19.9) 20 (29.0)
10-19 125 (33.7) 66 (29.2) 25 (36.2)
20+ 101 (27.2) 86 (38.1) 21 (30.4)
Missing 32 (8.6) 29 (12.8) 3 (4.4)
Cancer location Nipple, central portion,
upper/lower inner quadrant
84 (22.6) 43 (19.0) 16 (23.2)
Upper/lower outer quadrant 163 (43.9) 99 (43.8) 30 (43.5)
Overlapping lesion 84 (22.6) 47 (20.8) 16 (23.2)
Axillary tail/NOS 40 (10.8) 37 (16.4) 7 (10.1)
Surgery type
Partial/less than total
mastectomy/surgery, NOS
272 (73.3) 123 (54.4) 47 (68.1)
Total mastectomy/
modified radical mastectomy
99 (26.7) 103 (45.6) 22 (31.9)
Reconstructive surgery No 304 (81.9) 180 (79.7) 49 (71.0)
Yes 56 (15.1) 36 (15.9) 13 (18.8)
Missing 11 (3.0) 10 (4.4) 7 (10.1)
Number of excised lymph nodes 0 88 (23.7) 41 (18.1) 14 (20.3)
1–9
a
100 (27.0) 49 (21.7) 13 (18.8)
10+ 183 (49.3) 136 (60.2) 42 (60.9)
Radiation No 126 (34.0) 122 (54.0) 22 (31.9)
Yes 245 (66.0) 104 (46.0) 47 (68.1)
Chemotherapy No 250 (67.4) 124 (54.9) 39 (56.5)
Yes 121 (32.6) 102 (45.1) 30 (43.5)
Tamoxifen No 196 (52.8) 149 (65.9) 37 (53.6)
Yes 175 (47.2) 77 (34.1) 32 (46.4)
34
Table 1. Participant characteristics by race/ethnicity (continued)
Characteristics Categories
Non-Hispanic white
(N=371)
%
Black
(N=226)
%
Hispanic white
(N=69)
%
Sociodemograhpic factors
Marital status
Married 262 (70.6) 137 (60.6) 46 (66.7)
Widowed 12 (3.2) 17 (7.5) 5 (7.3)
Divorced/separated 67 (18.1) 57 (25.2) 16 (23.2)
Never married 30 (8.1) 15 (6.6) 2 (2.9)
Education High school or less 54 (14.6) 77 (34.1) 27 (39.1)
Some college 125 (33.7) 101 (44.7) 23 (33.3)
College graduate 86 (23.2) 27 (12.0) 7 (10.1)
Graduate studies 106 (28.6) 21 (9.3) 12 (17.4)
Insurance Yes 345 (93.0) 207 (91.6) 66 (95.7)
No 8 (2.2) 14 (6.2) 3 (4.4)
Missing 18 (4.9) 5 (2.2) 0 (0.0)
Health-related factors
Body mass index prior to diagnosis
(kg/m
2
)
<25 192 (51.8) 89 (39.4) 25 (36.2)
25-29.9 114 (30.7) 76 (33.6) 30 (43.5)
30+ 65 (17.5) 61 (27.0) 14 (20.3)
Menopausal status Premenopausal 143 (38.5) 89 (39.4) 27 (39.1)
Postmenopausal 196 (52.8) 111 (49.1) 34 (49.3)
Unknown 32 (8.6) 26 (11.5) 8 (11.6)
Hypertension No 292 (78.7) 119 (52.7) 57 (82.6)
Yes 79 (21.3) 107 (47.4) 12 (17.4)
Diabetes No 348 (93.8) 195 (86.3) 64 (92.8)
Yes 23 (6.2) 31 (13.7) 5 (7.3)
Arthritis No 271 (73.1) 164 (72.6) 45 (65.2)
Yes 100 (27.0) 62 (27.4) 24 (34.8)
Charlson Comorbidity Index 0 337 (90.8) 199 (88.1) 63 (91.3)
1–2 32 (8.6) 27 (12.0) 6 (8.70)
Missing 2 (0.5) 0 (0.0) 0 (0.00)
Hormonal factors
Oral contraceptives use prior to
diagnosis
No 86 (23.2) 81 (35.8) 22 (31.9)
Yes 285 (76.8) 145 (64.2) 47 (68.1)
Estrogen use prior to diagnosis
b
No 192 (51.8) 176 (77.9) 42 (60.9)
Yes 165 (44.5) 49 (21.7) 27 (39.1)
Missing 14 (3.8) 1 (0.4) 0 (0.0)
Progestin use prior to diagnosis
b
No 242 (65.2) 206 (91.2) 55 (79.7)
Yes 109 (29.4) 19 (8.4) 13 (18.8)
Missing 20 (5.4) 1 (0.4) 1 (1.5)
Lifestyle factors
Pack-years of smoking <100 cigarettes in lifetime,
<0.05 pack-years
183 (49.3) 110 (48.7) 33 (47.8)
0.05-5.3 58 (15.6) 38 (16.8) 16 (23.2)
5.4–20.5 55 (14.8) 45 (19.9) 11 (15.9)
>20.5 72 (19.4) 33 (14.6) 9 (13.0)
Missing 3 (0.8) 0 (0.0) 0 (0.0)
Alcohol intake year prior to diagnosis
(grams per day)
<1 137 (36.9) 170 (75.2) 22 (31.9)
1–6 58 (15.6) 29 (12.8) 12 (17.4)
>6 86 (23.2) 25 (11.1) 6 (8.7)
Missing 90 (24.3) 2 (0.9) 29 (42.0)
Sports and recreational activities year
prior to diagnosis (MET hours/week)
<0.5 75 (20.2) 129 (57.1) 19 (27.5)
0.5–20.0 192 (51.8) 71 (31.4) 35 (50.7)
>20.0 103 (27.8) 26 (11.5) 15 (21.7)
Missing 1 (0.3) 0 (0.0) 0 (0.0)
Abbreviations: MET: metabolic equivalent task; N: number; NOS: not otherwise specified
a
“1 to 9” category includes “at least one lymph node removed.”
b
Only postmenopausal women were included.
35
Table 2. Participant characteristics and their associations with lymphedema among 666 female
breast cancer survivors
Characteristics Categories N LE HR
a
95% CI HR
b
95% CI
Breast cancer and
treatment
Age at diagnosis (years) 35–44 123 43 1.00 1.00
45–49 128 43 0.98 0.64, 1.50 0.98 0.63, 1.52
50–54 175 46 0.80 0.52, 1.21 0.79 0.51, 1.23
55–59 128 35 0.83 0.53, 1.30 0.78 0.49, 1.23
60–64 112 23 0.59 0.35, 0.97 0.61 0.35, 1.06
Per unit of age 0.98 0.96, 1.00 0.98 0.96, 1.00
p trend=0.03 p trend=0.04
Disease stage In situ 155 15 0.26 0.15, 0.45
Localized 346 105 1.00
Regional 165 70 1.35 0.99, 1.85
Tumor size (mm) <10 178 36 1.00
10-19 216 66 1.54 1.02, 2.32
20+ 208 76 1.85 1.23, 2.77
Missing 64 12 0.80 0.41, 1.54
Per mm 1.01 1.00, 1.02
p trend=0.02
Cancer location Nipple, central portion,
upper/lower inner quadrant
143 44 1.15 0.80, 1.66
Upper/lower outer quadrant 292 83 1.00
Overlapping lesion 147 44 1.07 0.74, 1.55
Axillary tail, NOS 84 19 0.75 0.46, 1.24
Surgery type Partial/less than total
mastectomy/surgery, NOS
442 108 1.00 1.00
Total mastectomy/modified
radical mastectomy
224 82 1.52 1.13, 2.04 1.37 1.01, 1.85
Reconstructive surgery No 533 150 1.00
Yes 105 33 0.98 0.67, 1.45
Missing 28 7
Number of excised lymph
nodes
0 143 9 1.00 1.00
1–9
c
162 32 3.45 1.65, 7.23 2.80 1.32, 5.95
10+ 361 149 7.93 4.04, 15.56 6.05 3.01, 12.16
Per lymph node 1.06 1.04, 1.08 1.05 1.03, 1.07
p trend <0.001 p trend<0.001
Radiation No 270 73 1.00
Yes 396 117 1.21 0.90, 1.63
Chemotherapy No 413 83 1.00 1.00
Yes 253 107 2.28 1.69, 3.09 1.48 1.09, 2.02
Tamoxifen No 382 96 1.00
Yes 284 94 1.58 1.18, 2.12
Sociodemographic
factors
Race/ethnicity Non-Hispanic white 371 88 1.00 1.00
Black 226 82 1.62 1.20, 2.20 1.16 0.83, 1.61
Hispanic white 69 20 1.19 0.73, 1.94 1.04 0.63, 1.70
Marital status
Married 445 123 1.00
Widowed 34 10 1.25 0.64, 2.43
Divorced/separated 140 46 1.20 0.86, 1.69
Never married 47 11 0.85 0.45, 1.58
Education High school or less 158 48 1.00
Some college 249 75 0.92 0.63, 1.33
College graduate 120 39 1.14 0.73, 1.77
Graduate studies 139 28 0.65 0.40, 1.07
Insurance Yes 618 180 1.00
No 25 8 1.04 0.51, 2.12
Missing 23 2
Health-related factors
Body mass index prior to
diagnosis (kg/m
2
)
<25 306 70 1.00 1.00
25-29.9 220 67 1.40 0.99, 1.96 1.25 0.88, 1.76
30+ 140 53 1.83 1.27, 2.63 1.59 1.09, 2.31
Per unit of BMI 1.04 1.02, 1.07 1.04 1.01, 1.06
36
p trend<0.001 p trend=0.01
Menopausal status
d
Premenopausal 259 82 1.00
Postmenopausal 341 80 0.81 0.51, 1.28
Unknown 66 28 1.44 0.91, 2.26
Hypertension No 468 122 1.00 1.00
Yes 198 68 1.54 1.11, 2.12 1.49 1.06, 2.10
Diabetes No 607 174 1.00
Yes 59 16 0.96 0.57, 1.62
Arthritis No 480 135 1.00
Yes 186 55 1.19 0.86, 1.65
Charlson Comorbidity
Index
0 599 177 1.00
1–2 65 13 0.70 0.40, 1.24
Missing 2 0
Hormonal factors
Oral contraceptives use
prior to diagnosis
No 189 48 1.00
Yes 477 142 1.12 0.79, 1.58
Estrogen use prior to
diagnosis
d,e
No 120 26 1.00
Yes 210 52 1.36 0.83, 2.24
Missing 11 2
Progestin use prior to
diagnosis
d,e
No 205 52 1.00
Yes 119 23 0.94 0.56, 1.59
Missing 17 5
Lifestyle factors
Pack-years of smoking <100 cigarettes in lifetime,
<0.05 pack-years
326 92 1.00
0.05-5.3 pack-years 112 33 1.06 0.71, 1.58
5.4–20.5 pack-years 111 37 1.31 0.89, 1.93
>20.5 pack-years 114 27 0.93 0.60, 1.44
Missing 3 1
Per pack-year of smoking 1.00 0.99, 1.01
p trend=0.47
Alcohol intake year prior to
diagnosis (grams per day)
<1 329 102 1.00
1–6 99 25 0.82 0.52, 1.28
>6 117 29 0.94 0.61, 1.45
Missing 121 34 1.28 0.79, 2.09
Per gram of alcohol 0.99 0.97, 1.01
p trend=0.42
Sports and recreational
activities year prior to
diagnosis (MET
hours/week)
<0.5 223 68 1.00
0.5–20.0 298 85 1.13 0.80, 1.59
>20.0 144 37 1.09 0.71, 1.67
Missing 1 0
Per MET hour 1.00 0.99, 1.01
p trend=0.80
Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio; LE, lymphedema; MET, metabolic equivalent task;
N, number, NOS, not otherwise specified
a
Cox proportional hazard model adjusted for race/ethnicity (Non-Hispanic white (New Mexico), Non-Hispanic white (Western
Washington), black (Los Angeles), Hispanic (New Mexico)) and age at diagnosis (35-44, 45-49, 50-54, 55-59, 60-64). Age and
race/ethnicity were mutually adjusted.
b
Cox proportional hazard model included age, race/ethnicity, surgery type, number of excised lymph nodes, chemotherapy, body mass
index, and hypertension.
c
“1 to 9” category includes “at least one lymph node removed.”
d
Age in years (35 to 64) was used to adjust
for the age effect.
e
Only postmenopausal women were included.
37
Table 3. Participant characteristics and their associations with early onset and late onset lymphedema
Characteristics Categories
Early LE/
at risk
Late LE/
at risk
Early onset LE
HR (95% CI)
a
Late onset LE
HR (95% CI)
a
Breast cancer and treatment
Age at diagnosis
(years)
35–44 18/123 25/105 1.00 1.00 1.00 1.00
45–49 22/128 21/106 1.19 0.63, 2.24 0.83 0.46, 1.51
50–54 29/175 17/146 1.19 0.65, 2.17 0.51 0.27, 0.96
55–59 20/128 15/108 1.04 0.54, 2.00 0.59 0.31, 1.12
60–64 16/112 7/96 1.03 0.50, 2.10 0.31 0.13, 0.74
Per unit of age 1.00 0.97, 1.03 0.95 0.92, 0.98
p trend=0.91 P trend=0.002
Surgery type Partial/less than total mastectomy/
surgery, NOS
57/442 51/385 1.00 1.00 1.00 1.00
Total mastectomy/
modified radical mastectomy
48/224 34/176 1.41 0.95, 2.09 1.32 0.84, 2.06
Number of excised lymph nodes
0 5/143 4/138 1.00 1.00 1.00 1.00
1–9
c
19/162 13/143 2.83 1.05, 7.68 2.69 0.87, 8.32
10+ 81/361 68/280 4.96 1.97, 12.50 7.50 2.69, 20.88
Per lymph node 1.04 1.01, 1.06 1.06 1.03, 1.09
p trend=0.003 P trend<0.001
Chemotherapy No 47/413 36/366 1.00 1.00 1.00 1.00
Yes 58/253 49/195 1.28 0.86, 1.91 1.77 1.13, 2.78
Sociodemographic factors
Race/ethnicity White 50/371 38/321 1.00 1.00 1.00 1.00
Black 45/226 37/181 1.08 0.70, 1.65 1.26 0.79, 2.02
Hispanic 10/69 10/59 0.90 0.45, 1.78 1.24 0.61, 2.50
Health-related factors
Body mass index prior to diagnosis
(kg/m
2
)
<25 37/306 33/269 1.00 1.00 1.00 1.00
25-29.9 36/220 31/184 1.24 0.78, 1.98 1.25 0.76, 2.06
30+ 32/140 21/108 1.64 1.00, 2.67 1.52 0.87, 2.66
Per unit of BMI 1.03 1.00, 1.07 1.04 1.00, 1.08
p trend=0.05 P trend=0.07
Hypertension No 63/468 59/405 1.00 1.00 1.00 1.00
Yes 42/198 26/156 1.65 1.08, 2.54 1.30 0.80, 2.13
Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio; LE, lymphedema
a
Cox proportional hazards model with time-dependent covariate adjusted for all the other variables in the table.
b
p value for interaction was based on likelihood ratio test.
c
“1 to 9” category includes “at least one lymph node removed.”
38
Table 4. Risk factors for arm lymphedema stratified by race/ethnicity
Non-Hispanic white Black Hispanic white
Characteristics Categories N LE (%) HR
a
95% CI N LE (%) HR
a
95% CI N LE (%) HR
a
95% CI
Age at diagnosis
(years)
35-44 56 19 (34) 1.00 53 21 (40) 1.00 1.00 14 3 (21) 1.00 1.00
45-49 65 16 (25) 0.90 0.45, 1.77 46 20 (43) 1.00 0.53, 1.92 17 7 (41) 2.23 0.50, 10.07
50-54 112 28 (25) 0.77 0.42, 1.40 48 17 (35) 0.87 0.44, 1.72 15 1 (7) 0.40 0.04, 4.30
55-59 77 17 (22) 0.74 0.38, 1.44 37 13 (35) 0.69 0.33, 1.44 14 5 (36) 2.77 0.56, 13.77
60-64 61 8 (13) 0.42 0.17, 1.00 42 11 (26) 0.60 0.26, 1.34 9 4 (44) 9.17 1.43, 58.94
Surgery type
Partial/less than total
mastectomy/surgery, NOS
272 53 (19) 1.00 123 43 (35) 1.00 1.00 47 12 (26) 1.00 1.00
Total/modified radical
mastectomy
99 35 (35) 1.58 1.02, 2.45 103 39 (38) 1.15 0.72, 1.81 22 8 (36) 2.16 0.71, 6.54
Number of excised
lymph nodes
0 88 3 (3) 0.10 0.03, 0.32 41 5 (12) 0.41 0.16, 1.10 14 1 (7) 0.09 0.01, 0.80
1-9
b
100 18 (18) 0.48 0.28, 0.82 49 13 (27) 0.64 0.35, 1.18 13 1 (8) 0.09 0.01, 0.70
10+ 183 67 (37) 1.00 136 64 (47) 1.00 1.00 42 18 (43) 1.00 1.00
Chemotherapy
No 250 43 (17) 1.00 124 27 (22) 1.00 1.00 39 13 (33) 1.00 1.00
Yes 121 45 (37) 1.25 0.80, 1.95 102 55 (54) 2.69 1.61, 4.50 30 7 (23) 0.69 0.24, 2.03
Body mass index
(kg/m
2
)
<25 192 42 (22) 1.00 89 23 (26) 1.00 1.00 25 5 (20) 1.00 1.00
25-29.9 114 26 (23) 1.05 0.64, 1.73 76 32 (42) 1.58 0.90, 2.78 30 9 (30) 2.25 0.65, 7.72
30+ 65 20 (31) 1.50 0.86, 2.61 61 27 (44) 1.65 0.91, 2.98 14 6 (43) 3.94 0.99, 15.75
Hypertension
No 292 73 (25) 1.00 119 30 (25) 1.00 1.00 57 19 (33) 1.00 1.00
Yes 79 15 (19) 0.95 0.52, 1.74 107 52 (49) 2.73 1.65, 4.53 12 1 (8) 0.21 0.02, 2.00
Abbreviations: CI, confidence interval; HR, hazard ratio; LE, lymphedema; NOS, not otherwise specified
a
Cox proportional hazard model adjusted for all of the other variables in this table.
b
“1 to 9” category includes “at least one lymph node removed.”
39
Chapter 3
Symptoms of breast cancer-related lymphedema and quality of life among disease-
free female breast cancer survivors
Nearly three million women are living in the United States with a history of invasive
breast cancer (1). Although most women living with breast cancer have measured quality
of life (QOL) similar to that of unaffected women, some encounter specific problems that
decrease their QOL (2-5).
One of the common post-treatment problems encountered by breast cancer survivors is
arm lymphedema (6). Breast cancer treatment involving axillary lymph node dissection
or axillary radiotherapy can damage the lymphatic system which plays a crucial role in
the regulation of tissue fluid homeostasis, immune function, and fat metabolism (7). This
can lead to an abnormal accumulation of protein-rich fluid in the interstitial spaces of the
arm, resulting in swelling of the limb, i.e., lymphedema (8,9). Roughly 6% of women
who received sentinel lymph node biopsy and about 20% of women who received
axillary lymph node dissection develop lymphedema (10). Many of these lymphedema-
affected women experience physical discomfort, cosmetic deformity, and upper extremity
disability (11). As there is currently no cure for lymphedema and the condition is usually
chronic, many lymphedema-affected women face this potentially debilitating condition
continuously throughout the years of their survivorship (12).
Previous studies have consistently shown that breast cancer survivors with lymphedema
experienced poorer QOL than those without (13-15). However, most of these studies
40
focused on studying the presence or absence of lymphedema or arm/hand swelling in
relation to QOL; data on which symptom(s) of lymphedema negatively affect the QOL
remains scarce. As such information will help us design effective interventions to
maintain or improve the QOL of lymphedema-affected breast cancer survivors, we
examined the associations of number of lymphedema symptoms or specific symptoms of
lymphedema with various aspects of QOL among disease-free female breast cancer
survivors who participated in the Health, Eating, Activity, and Lifestyle (HEAL) Study.
MATERIALS AND METHODS
Study Setting, Subjects, and Recruitment
Details of the aims, study design, and recruitment procedures have been published
previously(16,17). Briefly, the HEAL Study is a multi-center, multi-ethnic prospective
cohort study consisting of 1,183 female breast cancer survivors. Women diagnosed with
first primary in situ or Stage I-IIIA invasive breast cancer between 1995 and 1999 were
identified and recruited into this study through Surveillance, Epidemiology, End Results
(SEER) registries in three regions of the United States: New Mexico, Los Angeles
County, and Western Washington. Women aged 40 to 65 years were recruited in Western
Washington; women aged 35 to 64 years were recruited in Los Angeles County; and
women under 92 years of age were recruited in New Mexico. We administered the first
assessment (baseline assessment) in person within the first year (on average, 6 months)
after a woman’s diagnosis; the second assessment, on average, 30 months after a
woman’s diagnosis (30-month assessment) via in-person interview or self-completed
41
questionnaire; and the third assessment, on average, 40 months after a woman’s diagnosis
(40-month assessment) by telephone interview or mailed questionnaire in New Mexico,
by mailed questionnaire plus telephone follow-up in Western Washington, and by
telephone interview in Los Angeles County.
In this analysis, we excluded 217 women who were younger than 35 years (n=4) or older
than 64 years (n=213) in order to provide similar age distributions across study sites. We
also excluded 14 women who did not receive any type of surgery; and 32 women whose
racial-ethnic classification was other than Hispanic, non-Hispanic White, or Black,
leaving 920 women. We further excluded 33 women who died before completing the 30-
month assessment, 48 women who had recurrence or new primary breast disease prior to
30-month assessment, 136 women who did not complete the 30-month assessment, 3
women who died before completing 40-month assessment, 12 women who had
recurrence or new primary breast disease between the 30-month assessment and the 40-
month assessment, 77 women who did not complete the 40-month assessment, and 11
women with incomplete or suspicious data on weight or height (n=6), employment status
(n=1), Charlson Comorbidity Index (n=1), or lymph node status (n=3). The final analytic
cohort consisted of 600 women (Figure 3).
We obtained informed consent from all participants at each assessment. The study was
approved by the institutional review boards of participating centers, in accord with
assurance filed with and approved by the U.S. Department of Health and Human
Services.
42
Data collection
Arm lymphedema. To determine the presence of lymphedema, we first presented our
definition of lymphedema to women during the interview at the 40-month assessment:
“Sometimes the arm on the side on which you had breast cancer becomes swollen
because of an accumulation of fluid in your arm. This is called lymphedema. Please do
not confuse this with the temporary swelling that occurs after surgery.” We then asked
the following “yes” or “no” question: “Have you experienced lymphedema in your arm at
any time since your breast cancer diagnosis?” For women who answered “yes,” we also
asked when they first experienced lymphedema symptoms (month and year), and whether
they were still experiencing lymphedema at the time of assessment.
Symptom reporting. We asked participants who self-reported lymphedema whether the
following symptoms were present in their arm(s) either at the initial occurrence of
lymphedema or at a later time: burning pain, numbness, feeling of tightness, feeling of
tension, feeling of heaviness, feeling of hardness, loss of mobility, raised temperature
(warmth)/redness, and dryness of skin.
Sociodemographic factors. We used information on race/ethnicity and educational status
collected at the baseline assessment and information on income, marital status, and
employment collected at the 30-month assessment.
Breast cancer-related and treatment-related factors. We obtained clinical data on breast
cancer diagnosis date, age at diagnosis, stage of cancer (SEER staging), treatment types
(surgeries, radiation therapy, chemotherapy, hormone therapy), and number of excised
43
lymph nodes through SEER cancer registry records and by abstracting hospital medical
records. Tamoxifen use at or prior to the baseline assessment was assessed using both
hospital medical records and self-report.
Health-related and hormone-related factors. Information on comorbid medical
conditions such as diabetes, hypertension, congestive heart failure, arthritis, and kidney
disease was collected through self-report and hospital records. The Charlson Comorbidity
Index score was calculated using information abstracted from hospital medical records
(18,19). The total Charlson Comorbidity Index was calculated without carcinoma.
Participants also reported information on height at age 18 years and weight five years
before diagnosis (black women in Los Angeles County) or weight one year before
diagnosis (Hispanic women in New Mexico and Non-Hispanic white women in New
Mexico and Western Washington) at the baseline assessment. Using these data, body
mass index (BMI) prior to diagnosis was calculated as weight (kg) divided by the square
of height (m
2
). Menopausal status was determined based on the following questionnaire
data: age, date of last menstruation, and hysterectomy and oophorectomy status.
Lifestyle factors. The baseline assessment also captured information on sports and
recreational physical activity in the year prior to diagnosis, history of smoking, and
alcohol intake in the year prior to diagnosis. Physical activity was measured using the
Modifiable Activity Questionnaire which has been previously validated (20). We
calculated MET hours per week of sports and recreational physical activity (21), pack-
years of smoking, and grams of alcohol per day.
44
Quality of life
Health-related quality of life. We used the Medical Outcomes Study 36-item short form
health survey (MOS SF-36) to assess participants’ health-related QOL (HRQOL). This
widely-used measure consists of 36 items that assess eight health concepts: 1) limitations
in physical activities due to health problems (Physical Functioning); 2) limitations in
usual role activities due to physical health problems (Role-Physical); 3) bodily pain; 4)
general health perceptions (General Health); 5) energy and fatigue (Vitality); 6)
limitations in social activities due to physical or emotional problems (Social
Functioning); 7) limitations in usual role activities due to emotional problems (Role-
Emotional); and 8) psychological distress and well-being (Mental Health) (22). SF-36
subscales were scored in reference to the 1998 general US population with a transformed
mean of 50 and a standard deviation of 10. These eight scales were then used to calculate
physical component summary (PCS) and mental component summary (MCS) scales. The
scores range from 0 to 100 with higher scores indicating higher levels of functioning and
health.
Perceived stress. We used a four-item version of the Perceived Stress Scale to measure
the degree to which situations in one’s life are appraised as stressful (23). Questions
asked: how often have you felt you were “unable to control important things in your life,”
“confident about your ability to handle personal problems,” “things were going your
way,” and “difficulties were piling up so high that you could not overcome them” during
the last 4 weeks. Items were rated on 5-point likert response scales and were coded to
range from 0 to 4. For each participant, all four items were summed to obtain a total
45
score ranging from 0 to 16 with higher score suggesting a greater degree of stress.
Fear of recurrence. We used a 5-item version of the Fear of Recurrence Scale (24) which
has not been validated. Participants were asked to indicate how much they agree or
disagree to statements that express concerns about future health status; for example, “I
would like to feel more certain about my health” and “I worry that my cancer will
return.” Participants rated their responses on a 5-point likert scale from 1 (strongly
disagree) to 5 (strongly agree). The total score ranges from 5 to 25 with higher score
suggesting greater fear of recurrence.
Sexual functioning. We also used Cancer Rehabilitation Evaluation System (CARES)
Sexual Functioning Summary Scale to measure sexual health after diagnosis (25).
Participants who were sexually active during the last six months of interview rated “Lack
of a sexual interest”, “Difficulty in becoming sexually aroused”, “Unable to relax and
enjoy sex”, and “Difficulty in having an orgasm” on a 4-point likert response scale from
0 (not a problem) to 3 (a serious problem). The severity score ranges from 0 to 12 with
higher scores indicating more difficulty or problems in sexual health.
Statistical methods
The Chi-square tests were used to compare socio-demographic factors, clinical
characteristics, breast cancer treatment-related factors, menopausal status, BMI, and
comorbidity across the three groups of women: no lymphedema, 0 to 2 lymphedema
symptoms, and 3 or more lymphedema symptoms. We used ANCOVA to compare each
of the following outcome measures across these three groups: SF-36 PCS and MCS
46
Scales, eight SF-36 subscales, Perceived Stress Scale, Fear of Recurrence Scale, and
Sexual Functioning Summary Scale. Due to non-normal distribution of the data, we used
rank-based ANCOVA for the analysis of SF-36 subscales, Fear of Recurrence, and
Sexual Functioning Summary Scale. Scheffé's post-hoc pair-wise comparisons were
performed (26). The following factors were considered for adjustment: race/ethnicity
(non-Hispanic white women; black women; and Hispanic white women), age at the 40-
month assessment (years), disease stage at diagnosis (in situ, localized, regional), breast
cancer treatment-related factors [chemotherapy (yes, no), radiation (yes, no), tamoxifen
(yes, no), number of excised lymph nodes (0-9, 10-19, 20 or above), surgery type (less
than total mastectomy, total mastectomy/radical mastectomy)], marital status at the 30-
month assessment (married or living with a partner, widowed or divorced or separated,
never married), education (high school or less, some college, college degree, graduate
studies), employment status at the 30-month assessment (currently working,
unemployed/not working/retired/disabled), income at the 30-month assessment ($50,000
or less, more than $50,000, missing), menopausal status (premenopausal,
postmenopausal, unknown), BMI (kg/m
2
) prior to diagnosis (less than 25, 25 to 29.9, 30
or above), Charlson Comorbidity Index (0, 1 to 2), and time since diagnosis (months).
We chose age at the 40-month assessment, race/ethnicity, disease stage, and BMI for
adjustment since these factors altered the effect estimates by 15% or more for PCS or
MCS.
Subsequent analyses focused on examining specific lymphedema symptoms in relation to
HRQOL measured by MOS SF-36, and the Perceived Stress Scale. By fitting ANCOVA,
47
we compared PCS, MCS, and Perceived Stress scores across three groups of women:
women without lymphedema, women with a particular lymphedema symptom, and
women without the particular symptom. Scheffe’s method was used for pair-wise
comparisons. All models were adjusted for age at the 40-month assessment,
race/ethnicity, disease stage, and BMI. We then mutually adjusted for the symptoms that
differentiated between women with a specific symptom and women without by including
all of those symptoms in a model simultaneously. This analysis was performed only
among women who reported lymphedema. For the symptoms that were independently
associated with PCS or MCS, we further examined the associations between the
symptoms and each of the subscales of SF-36.
Lastly, we compared the frequencies of self-reported symptoms that had a significant
impact on HRQOL across racial/ethnic groups and age groups (38 to 49, 50 to 59, 60 to
69 years) using the Chi-square test.
All analyses were performed using SAS software (SAS version 9.2; SAS institute Inc,
Cary, NC).
RESULTS
The study population consisted of three racial/ethnic groups: 350 non-Hispanic white
women [New Mexico (35%), Western Washington (23%)], 192 black women [Los
Angeles (32%)], and 58 Hispanic women [New Mexico (10%)]. About 23% of the
women had in situ breast cancer, 54% had localized breast disease, and 23% had regional
breast disease. Four hundred two women (67%) were working at the time of the 30-
48
month assessment. A total of 136 (23%) women indicated that they experienced
lymphedema prior to the 40-month assessment; 99 (73%) of those women were
experiencing lymphedema at the time of the 40-month assessment. The mean number of
symptoms was 2.9 [standard deviation (SD): 2.5]. The most common self-reported
symptoms were: feeling of heaviness (52%), numbness (49%), and feeling of tightness
(48%). A total of 73 (54%) women received some type of treatment for lymphedema
such as compression by elastic sleeve (66%), arm elevation (62%), and massage therapy
(59%). Frequencies of participant characteristics by groups defined by presence of
lymphedema and the number of self-reported lymphedema symptoms (no lymphedema, 0
to 2 symptoms, and 3 or more symptoms) are summarized in Table 5. In this study
cohort, race/ethnicity, age, disease stage, surgery type, number of excised lymph nodes,
chemotherapy, tamoxifen, and BMI were significantly associated with presence of
lymphedema/number of self-reported lymphedema symptoms.
Presence of lymphedema/number of lymphedema symptoms and quality of life
Women with three or more lymphedema symptoms had significantly lower HRQOL (SF-
36 PCS and all SF-36 subscales with the exception of Vitality, Role-Emotional, and
Mental Health) than women without lymphedema (Table 6). We also found that QOL
measured by SF-36 PCS, SF-36 Physical Functioning, and Perceived Stress Scale was
worse among women with three or more symptoms as compared to lymphedema-affected
women who reported 0 to 2 symptoms (Table 6). A total of 352 (59%) women indicated
that they were sexually active in the past six months of 40-month assessment, and 347
(99%) of the 352 women completed the CARES Sexual Functioning Summary Scale.
49
No significant difference in Sexual Functioning Summary Scale was observed across the
three groups defined by presence of lymphedema/the number of lymphedema symptoms
(P=0.68). We also did not observe a significant difference in SF-36 MCS (P=0.22) or
Fear of Recurrence (P=0.22) across the three groups (Table 6).
Specific-lymphedema symptoms and health-related quality of life
SF-36 Physical Component Summary. We present adjusted means of PCS by groups
defined by presence of lymphedema/lymphedema symptom (no lymphedema,
lymphedema without a specific symptom, lymphedema with a specific symptom) in
Table 7. For all of the nine symptoms asked in this study, PCS scores differed across the
groups of women without lymphedema, women without a particular symptom, and
women with a particular symptom after adjusting for age, race/ethnicity, disease stage,
and BMI (Table 7). In an analysis comparing each pair, we found that women who
reported numbness, tension, loss of mobility, or raised temperature/redness in the
lymphedema-affected arm had a lower PCS than women who did not report such
symptoms (Table 7). When these symptoms were mutually adjusted, only feeling of
tension was independently associated with PCS (P=0.003). More specifically, women
who were feeling tension in the affected-arm had a significantly lower HRQOL in the
following domains: Physical Functioning (P<0.001), Role-Physical (P=0.002), Bodily
Pain (P=0.001), General Health (P=0.03), and Social Functioning (P=0.008), and Mental
Health (P=0.02) (data not shown).
SF-36 Mental Component Summary. We present adjusted means of MCS by groups
50
defined by presence of lymphedema/lymphedema symptom (no lymphedema,
lymphedema without burning pain, lymphedema with burning pain) in Table 8. We found
a statistically significant difference in MCS between women with and without burning
pain in the lymphedema-affected arm after adjusting for age, race/ethnicity, disease stage,
and BMI (P=0.02) (Table 8). More specifically, women who were experiencing burning
pain in the affected-arm had a significantly lower QOL in the following domains: Role-
Physical (P=0.04), Social Functioning (P=0.03), and Role-Emotional (P=0.02) (data not
shown).
Perceived Stress. We present adjusted means of Perceived Stress Scale by groups defined
by presence of lymphedema/lymphedema symptom (no lymphedema, lymphedema
without a specific symptom, lymphedema with a specific symptom) in Table 9. Women
with burning pain, feeling of tension, or raised temperature/redness perceived a higher
level of stress compared to women without such symptoms (Table 9). When these three
symptoms were included in the model simultaneously, none of the symptoms remained
significant (P>0.05); however, when raised temperature/redness was dropped from the
model, the associations with both burning pain (P=0.04) and feeling of tension (P=0.02)
were statistically significant (data not shown).
Finally, we present the data on distributions of women who reported tension and burning
pain across race/ethnicity and age groups in Table 10. Black women and non-Hispanic
white women were more likely to report a feeling of tension than Hispanic white women
(P=0.03).
51
DISCUSSION
In this multi-racial-ethnic cohort of disease-free female breast cancer survivors, 136
indicated that they experienced lymphedema. These women reported various symptoms
in their lymphedema-affected arm with the most common symptoms being a feeling of
heaviness, numbness, and a feeling of tightness. Our study demonstrated negative
associations between self-reported lymphedema symptoms and QOL. In particular,
women who indicated feeling of tension had decreased physical HRQOL and women
who indicated burning pain had decreased mental HRQOL. These symptoms were also
associated with a greater level of perceived stress.
Our study agrees with the literature that women with lymphedema have lower physical
HRQOL than women without lymphedema (13,15,27). In our study, women who
reported 3 or more symptoms had significantly lower physical HRQOL than women who
reported less than 3 symptoms. Among the 9 symptoms evaluated in this study, a feeling
of tension in the lymphedema-affected arm had the strongest association with decreased
physical HRQOL. We also found that a feeling of tension in the arm negatively affected
Physical Functioning, Role-Physical, Bodily Pain, General Health, and Social
Functioning, and Mental Health.
Although the associations of lymphedema symptoms with mental HRQOL were not as
strong as those with physical HRQOL, we observed a significant association between
lymphedema-related burning pain and mental HRQOL. More specifically, having
burning pain in the arm increased limitations in usual role activities due to physical or
52
emotional health problems, and limitations in social activities due to physical or
emotional problems. Wilson et al (28) used the SF-36 and FLIC to study the association
between lymphedema and HRQOL and found that FLIC yielded a greater discriminative
power compared to SF-36 in detecting the effect of lymphedema on mental HRQOL. It is
possible that our non-significant results could have been due to lack of statistical power
or discriminative power. Our results are also consistent with other studies that examined
the association between pain and QOL (27,29). Dawes’ study (29) found strong
relationships between pain and activity limitation, participation restriction, and sub-
optimal HRQOL after breast cancer surgery. Another study by Hormes et al (27) also
demonstrated a strong relationship between pain and decreased QOL regardless of arm
swelling. These data highlight the importance of measuring and reducing pain among
lymphedema-affected breast cancer survivors.
Our analysis also showed that burning pain and a feeling of tension increased the levels
of perceived stress. As burning pain and a feeling of tension were also associated with
HRQOL, this finding emphasizes the importance of paying special attention to
lymphedema-affected women with these symptoms.
Fear of cancer recurrence is one of the most important sources of distress after breast
cancer treatment (30). While qualitative studies have shown that the lymphedema-
induced swelling reminds the survivors of cancer and increases their fear of recurrence
(31,32), our quantitative analysis failed to demonstrate an association between number of
lymphedema symptoms and the levels of fear of recurrence.
53
Passik et al (33) found that women with upper extremity lymphedema, particularly in
their dominant hand, had higher levels of sexual dysfunction. Hormes et al (27) and
Newman et al (34) also found an association between lymphedema and sexual
functioning, while Dawes et al (29) found no statistically significant difference in sexual
functioning or sexual enjoyment between women with lymphedema and the population
norm. Our study failed to demonstrate an association between number of lymphedema
symptoms and sexual functioning; however, we found that women who reported more
lymphedema symptoms were more likely to be sexually inactive after multivariable
adjustment (P=0.048) (data not shown). Absence of sexual activity may indicate presence
of sexual dysfunction; therefore, our null results should not discourage researchers and
clinical practitioners from improving this area of QOL in lymphedema-affected breast
cancer survivors.
Ridner et al (35) found that women previously treated for lymphedema continued to have
lymphedema symptoms including alteration in limb sensation. Similarly, our study
showed that the number of symptoms did not differ significantly by treatment status
(prior treatment, no prior treatment), which may indicate that their lymphedema
symptoms remained after treatment (data not shown). Treatment of lymphedema has been
primarily focused on reducing the size/volume of the lymphedema-affected arm (29,36).
Thus it is important to encourage treatment to manage not only swelling, but also other
symptoms such as burning pain and tension in the affected arm using evidence-based
measures. A randomized study conducted by Johansson et al (37) showed that the manual
lymphatic drainage with or without compression bandaging reduced the feeling of tension
54
in the lymphedema-affected breast cancer survivors. In Hamner’s study (38), complete
decongestive therapy reduced both volume and pain among lymphedema patients. More
studies are needed to evaluate the effects of these treatments on different symptoms of
lymphedema as well as the QOL of breast cancer survivors with lymphedema.
The following limitations should be considered when interpreting the results. In the
HEAL Study, lymphedema was determined based on self-report; therefore
misclassification of lymphedema may be possible. Although confusion with temporary
swelling may cause false-positives and decreased specificity, women in the HEAL Study
were told not to confuse with temporary swelling caused by surgery. Furthermore, the
questionnaire was completed, on average, 40 months after breast cancer diagnosis, thus
misclassification due to confusion with temporary swelling is unlikely. A potential lack
of discriminative power of SF-36 MCS should also be considered when examining the
associations between lymphedema symptoms and mental HRQOL among breast cancer
survivors. Another limitation is that QOL was measured only once at the 40-month
assessment, thus we were unable to evaluate the change in the levels of QOL.
55
In conclusion, lymphedema-related burning pain and feeling of tension independently
decreased HRQOL and increased the level of perceived stress in this cohort of disease-
free female breast cancer survivors. This study suggests the importance of continuous
treatment for lymphedema by managing not only swelling, but also tension and burning
pain in the lymphedema-affected arm in order to improve the QOL of lymphedema-
affected breast cancer survivors. Further studies are needed to provide information on
effective measures to manage these symptoms and the effects of those measures on the
QOL.
56
References
1. American Cancer Society. Breast Cancer Facts & Figures 2013-2014. 2013;:1–40.
2. Mols F, Vingerhoets AJJM, Coebergh JW, et al. Quality of life among long-term
breast cancer survivors: A systematic review. Eur J Cancer. 2005;41(17):2613–
2619.
3. Helgeson VS, Tomich PL. Surviving cancer: a comparison of 5-year disease-free
breast cancer survivors with healthy women. Psycho-Oncology. 2005;14(4):307–
317.
4. Hsu T, Ennis M, Hood N, et al. Quality of life in long-term breast cancer
survivors. J. Clin. Oncol. 2013;31(28):3540–3548.
5. Dorval MM, Maunsell EE, Deschênes LL, et al. Long-term quality of life after
breast cancer: comparison of 8-year survivors with population controls. J. Clin.
Oncol. 1998;16(2):487–494.
6. Erickson VS, Pearson ML, Ganz PA, et al. Arm edema in breast cancer patients. J.
Natl. Cancer Inst. 2001;93(2):96–111.
7. Norrmen C, Tammela T, Petrova TV, et al. Biological Basis of Therapeutic
Lymphangiogenesis. Circulation. 2011;123(12):1335–1351.
8. Rockson SG. Lymphedema. Am J Med. 2001;110(4):288–295.
9. Sakorafas GH, Peros G, Cataliotti L, et al. Lymphedema following axillary lymph
node dissection for breast cancer. Surgical Oncology. 2006;15(3):153–165.
10. DiSipio T, Rye S, Newman B, et al. Incidence of unilateral arm lymphoedema
after breast cancer: a systematic review and meta-analysis. The Lancet Oncology.
2013;14(6):500–515.
11. Petrek JA, Pressman PI, Smith RA. Lymphedema: current issues in research and
management. CA-ATLANTA. 2000;
12. Petrek JA, Senie RT, Peters M, et al. Lymphedema in a cohort of breast carcinoma
survivors 20 years after diagnosis. Cancer. 2001;92(6):1368–1377.
13. Ahmed RL, Prizment A, Lazovich D, et al. Lymphedema and Quality of Life in
Breast Cancer Survivors: The Iowa Women's Health Study. Journal of Clinical
Oncology. 2008;26(35):5689–5696.
14. Beaulac SM, McNair LA, Scott TE, et al. Lymphedema and quality of life in
survivors of early-stage breast cancer. Arch Surg. 2002;137(11):1253–1257.
57
15. Kwan W. Chronic Arm Morbidity After Curative Breast Cancer Treatment:
Prevalence and Impact on Quality of Life. Journal of Clinical Oncology.
2002;20(20):4242–4248.
16. Irwin ML, Crumley D, McTiernan A, et al. Physical activity levels before and after
a diagnosis of breast carcinoma. Cancer. 2003;97(7):1746–1757.
17. McTiernan A. Adiposity and Sex Hormones in Postmenopausal Breast Cancer
Survivors. Journal of Clinical Oncology. 2003;21(10):1961–1966.
18. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic
comorbidity in longitudinal studies: development and validation. J Chronic Dis.
1987;40(5):373–383.
19. Harlan LC, Klabunde CN, Ambs AH, et al. Comorbidities, therapy, and newly
diagnosed conditions for women with early stage breast cancer. J Cancer Surviv.
2009;3(2):89–98.
20. Kriska AM, Caspersen CJ. Introduction to the collection of physical activity
questionnaires. In A Collection of Physical Activity Questionnaires for Health-
Related Research. Medicine & Science in Sports & Exercise. 1997;29(6
Suppl):S5–S9 and S73–S78.
21. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities:
an update of activity codes and MET intensities. Medicine & Science in Sports &
Exercise. 2000;32(9 Suppl):S498–504.
22. Ware JEJ, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I.
Conceptual framework and item selection. Med Care. 1992;30(6):473–483.
23. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J
Health Soc Behav. 1983;24(4):385–396.
24. Northouse LL. Mastectomy patients and the fear of cancer recurrence. Cancer
Nurs. 1981;4(3):213–220.
25. Schag CA, Heinrich RL, Aadland RL, et al. Assessing problems of cancer patients:
psychometric properties of the cancer inventory of problem situations. Health
Psychol. 1990;9(1):83–102.
26. Scheffé H. The Analysis of Variance. John Wiley & Sons; 1999 1 p.
27. Hormes JM, Bryan C, Lytle LA, et al. Impact of lymphedema and arm symptoms
on quality of life in breast cancer survivors. Lymphology. 2010;43(1):1–13.
28. Wilson RWR, Hutson LML, Vanstry DD. Comparison of 2 quality-of-life
questionnaires in women treated for breast cancer: the RAND 36-Item Health
Survey and the Functional Living Index-Cancer. Phys Ther. 2005;85(9):851–
58
860.
29. Dawes DJ, Meterissian S, Goldberg M, et al. Impact of lymphoedema on arm
function and health-related quality of life in women following breast cancer
surgery. J Rehabil Med. 2008;40(8):651–658.
30. Kuehn T, Klauss W, Darsow M, et al. Long-term morbidity following axillary
dissection in breast cancer patients--clinical assessment, significance for life
quality and the impact of demographic, oncologic and therapeutic factors. Breast
Cancer Res Treat. 2000;64(3):275–286.
31. Fu MR, Rosedale M. Breast Cancer Survivors' Experiences of Lymphedema-
Related Symptoms. J Pain Symptom Manage. 2009;38(6):849–859.
32. Rosedale M, Fu MR. Confronting the unexpected: temporal, situational, and
attributive dimensions of distressing symptom experience for breast cancer
survivors. Oncology Nursing Forum. 2010;37(1):E28–33.
33. Passik SD, Newman ML, Brennan M, et al. Predictors of psychological distress,
sexual dysfunction and physical functioning among women with upper extremity
lymphedema related to breast cancer. Psycho-Oncology. 1995;4(4):255–263.
34. Newman ML, Brennan M, Passik S. Lymphedema complicated by pain and
psychological distress: a case with complex treatment needs. J Pain Symptom
Manage. 1996;12(6):376–379.
35. Ridner SH. Quality of life and a symptom cluster associated with breast cancer
treatment-related lymphedema. Support Care Cancer. 2005;13(11):904–911.
36. Moseley A, Carati C, Piller N. A systematic review of common conservative
therapies for arm lymphoedema secondary to breast cancer treatment. Annals of
Oncology. 2006;18(4):639–646.
37. Johansson K, Albertsson M, Ingvar C, et al. Effects of compression bandaging
with or without manual lymph drainage treatment in patients with postoperative
arm lymphedema. Lymphology. 1999;32(3):103–110.
38. Hamner JB, Fleming MD. Lymphedema therapy reduces the volume of edema and
pain in patients with breast cancer. Ann Surg Oncol. 2007;14(6):1904–1908.
59
Figure 3. Recruitment flow and data collection
Data collection: Timing and variables
Recruitment flow
Baseline assessment
(on average, 6 months following breast cancer
diagnosis)
- Sociodemographics: race/ethnicity, education
- Body mass index
- Physical activity
Hospital medical records (baseline data)/SEER
cancer registry records
- Date of birth
- Breast cancer stage, tumor size and location
- Treatment (surgery, number of excised lymph
nodes, radiation, chemotherapy, tamoxifen)
- Comorbidity (Charlson Comorbidity Index)
1,183 participants completed the baseline
assessment.
- 615 New Mexico
- 202 Western Washington
- 366 Los Angeles County
920 of the 1,183 participants who completed
the baseline assessment met the criteria for the
analysis.
Exclusions, n (reason):
- 4 (age <35 years)
- 213 (age>65 years)
- 14 (no surgery)
- 32 (race is other than
non-Hispanic white, black, and
Hispanic white)
703 of the 920 participants who were eligible
for the analysis completed the second
assessment.
Exclusions, n (reason):
- 33 (deceased)
- 48 (recurrence or new breast cancer)
- 136 (did not complete the second
assessment)
Second assessment
(on average, 30 months following breast cancer
diagnosis)
- Sociodemographics: marital status,
employment status, income
Third assessment
(on average, 40 months following breast cancer
diagnosis)
- Self-reported symptoms of lymphedema
- Self-reported treatment of lymphedema
- Health-related quality of life (MOS SF-36)
- Other QOL measures: Perceived Stress Scale,
Fear of Recurrence, Sexual Functioning
Summary Scale
611 of the 703 participants who completed the
second assessment completed the third
assessment.
Exclusions, n (reason):
- 3 (deceased)
- 12 (recurrence or new primary breast
cancer)
- 77 (did not complete the third
assessment)
600 of the 611 participants who completed the
third assessment comprised the analytic
cohort.
Exclusions, n (reason):
- 6 (incomplete information on
weight or height)
- 1 (incomplete information on CCI)
- 1 (missing information on work
status)
- 3 (suspicious information on
number of excised lymph nodes)
60
Table 5. Characteristics of 600 women by number of self-reported lymphedema symptoms
Characteristics
No lymphedema
N (%)
Lymphedema with
0 to 2 symptoms
N (%)
Lymphedema with
3+ symptoms
N (%)
P value
Race/ethnicity
Non-Hispanic White 285 (81) 36 (10) 29 (8)
Black 136 (71) 13 (7) 43 (22)
Hispanic white 43 (74) 10 (17) 5 (9) <0.001
Age at 40-month assessment (years)
35-49 100 (69) 15 (10) 29 (20)
50-59 212 (78) 28 (10) 33 (12)
60-69 152 (83) 16 (9) 15 (8) 0.02
Disease stage
In situ 133 (96) 2 (1) 4 (3)
Localized 240 (74) 40 (12) 44 (14)
Regional 91 (66) 17 (12) 29 (21) <0.001
Surgery type
Partial/less than total
mastectomy/surgery, NOS
317 (81) 41 (10) 36 (9)
Total mastectomy/
modified radical mastectomy
147 (71) 18 (9) 41 (20) 0.001
Number of excised lymph nodes
0-9
a
251 (91) 11 (4) 13 (5)
10-19 164 (66) 37 (15) 48 (19)
20 or above 49 (64) 11 (14) 16 (21) <0.001
Radiation
No 189 (78) 19 (8) 35 (14)
Yes 275 (77) 40 (11) 42 (12) 0.29
Chemotherapy
No 316 (84) 28 (7) 34 (9)
Yes 148 (67) 31 (14) 43 (19) <0.001
Tamoxifen
Never 232 (80) 25 (9) 32 (11)
Prior use 94 (72) 7 (5) 30 (23)
Current use at 40-month assessment 138 (77) 27 (15) 15 (8) <0.001
Marital status at 30-month assessment
Married or living with a partner 286 (79) 34 (9) 40 (11)
Widowed/divorced/separated 146 (74) 22 (11) 29 (15)
Never married 32 (74) 3 (7) 8 (19) 0.44
b
Education
High school or less 109 (77) 14 (10) 18 (13)
Some college 172 (76) 24 (11) 29 (13)
College graduate 81 (72) 10 (9) 22 (19)
Graduate school 102 (84) 11 (9) 8 (7) 0.17
Employment status at 30-month
assessment
Currently working 314 (78) 43 (11) 45 (11)
Unemployed/not working/retired/disabled 150 (76) 16 (8) 32 (16) 0.17
Income at 30-month assessment
$50,000 or less 229 (77) 24 (8) 45 (15)
61
More than $50,000 207 (77) 31 (12) 30 (11)
Missing 28 (82) 4 (12) 2 (6) 0.28
Menopausal status at 30-month
assessment
Premenopausal 101 (79) 10 (8) 17 (13)
Postmenopausal 339 (78) 44 (10) 51 (12)
Unknown 24 (63) 5 (13) 9 (24) 0.20
Body mass index (kg/m
2
) prior to
diagnosis
Less than 25 225 (81) 21 (8) 31 (11)
25-29.9 153 (78) 26 (13) 17 (9)
30 or above 86 (68) 12 (10) 29 (23) 0.001
Charlson Comorbidity Index
0 413 (77) 58 (11) 68 (13)
1-2 51 (84) 1 (2) 9 (15) 0.08
Abbreviation: MET=metabolic equivalent
a
“1 to 9” category includes “at least one lymph node removed.”
b
Women with missing values were excluded.
62
Table 6. Comparison of quality of life measures by number of self-reported lymphedema symptoms
Quality of life measures
No lymphedema
Lymphedema with
0 to 2 self-reported symptoms
Lymphedema with
3+ self-reported symptoms
Mean
(SE)/
Median (IQR)
Mean (SE)/
Median (IQR)
Mean (SE)/
Median (IQR)
P value
a
MOS SF-36 PCS 49.2 (0.6) 47.3 (1.3) 43.1 (1.2)
c,d
<0.001
MOS SF-36 MCS 48.6 (0.6) 48.7 (1.5) 46.3 (1.3) 0.22
MOS SF-36 Subscale
b
Physical Functioning 50.2 (45.5-54.8) 50.2 (40.8-54.8) 40.8 (31.5-50.2)
c,d
<0.001
Role-Physical 56.2 (42.1-56.2) 56.2 (35.0-56.2) 42.1 (28.0-56.2)
c
<0.001
Bodily Pain 51.6 (42.2-62.7) 51.6 (37.5-55.9) 46.5 (37.5-55.9)
c
<0.001
General Health 53.2 (46.2-57.9) 52.3 (43.9-57.9) 48.5 (40.6-55.6)
c
0.005
Vitality 51.4 (44.3-58.5) 51.4 (44.3-58.5) 49.1 (39.6-53.8) 0.06
Social Functioning 51.7 (46.3-57.1) 51.7 (40.9-57.1) 46.3 (35.4-57.1)
c
0.02
Role-Emotional 55.3 (44.8-55.3) 55.3 (34.3-55.3) 55.3 (34.3-55.3) 0.06
Mental Health 52.7 (43.6-57.3) 52.7 (45.9-57.3) 50.4 (43.6-57.3) 0.10
Perceived Stress Scale 8.9 (0.2) 8.4 (0.4) 9.7 (0.4)
d
0.03
Fear of Recurrence Scale
b
16 (13-19) 17 (14-20) 17 (14-19) 0.22
Sexual Functioning Summary Scale
b,e
3 (1-6) 2 (0-7) 3 (1-6) 0.68
Abbreviation: IQR, interquartile range; MCS, mental component summary; MOS, medical outcomes study;
PCS, physical component summary; SE, standard error; SF, Short form
a
Adjusted for age at 40-month assessment, race/ethnicity, disease stage, and body mass index
b
Rank-based multivariable analysis of variance was used. The median and IQR were presented
c
Significantly different from women without lymphedema symptoms based on Scheffé's pair-wise multiple comparison test
d
Significantly different from women with 0 to 2 lymphedema symptoms based on Scheffé's pair-wise multiple comparison test
e
Only 347 women who were sexually active in the past 6 months of 3-year assessment were included in the analysis
63
Table 7. The associations between self-reported lymphedema symptoms and physical health-related
quality of life score
MOS SF-36 PCS
Total
N
Below 50
N (%)
Adjusted
mean
a
Adjusted
mean
a,b
P value
a,b
No lymphedema 464 186 (40) 49.6
Lymphedema 136 89 (65) 45.3
c
No lymphedema 464 186 (40) 49.2
Lymphedema, no burning pain 113 72 (64) 45.5
c
Lymphedema with burning pain 23 17 (74) 41.9
c
No lymphedema 464 186 (40) 49.2
Lymphedema, no numbness 70 36 (51) 47.2 44.6
Lymphedema with numbness 66 53 (80) 42.6
c,d
41.8 0.11
No lymphedema 464 186 (40) 49.2
Lymphedema, no tightness 71 39 (55) 46.5
Lymphedema with tightness 65 50 (77) 43.3
c
No lymphedema 464 186 (40) 49.2
Lymphedema, no tension 100 58 (58) 47.0
c
46.4
Lymphedema with tension 36 31 (86) 39.0
c,d
40.0 0.003
No lymphedema 464 186 (40) 49.2
Lymphedema, no heaviness 65 37 (57) 46.7
Lymphedema with heaviness 71 52 (73) 43.2
c
No lymphedema 464 186 (40) 49.3
Lymphedema, no hardness 103 62 (60) 46.0
c
Lymphedema with hardness 33 27 (82) 41.8
c
No lymphedema 464 186 (40) 49.3
Lymphedema, no loss of mobility 105 66 (63) 46.1
c
43.7
Lymphedema with loss of mobility 31 23 (74) 41.1
c,d
42.7 0.66
No lymphedema 464 186 (40) 49.2
Lymphedema, no raised temperature/redness 100 58 (58) 46.2
c
43.9
Lymphedema with raised temperature/redness 36 31 (86) 41.2
c,d
42.6 0.52
No lymphedema 464 186 (40) 49.2
Lymphedema, no dry skin 98 61 (62) 45.6
c
Lymphedema with dry skin 38 28 (74) 43.4
c
Abbreviation: MOS, medical outcomes study; PCS, physical component summary; SF, short form
a
Adjusted for age at 40-month assessment, race/ethnicity, disease stage, body mass index
b
Mutually adjusted for the following symptoms: numbness, tension, loss of mobility, and raised temperature/redness (analysis of 136
lymphedema-affected women)
c
Significantly different from women without lymphedema based on Scheffé’s pair-wise multiple comparison test
d
Significantly different from lymphedema-affected women without a particular symptom based on Scheffé’s pair-wise multiple
comparison test
64
Table 8. The associations between self-reported lymphedema symptoms and mental health-related
quality of life score
MOS SF-36 MCS Total N
Below 50
N (%)
Adjusted
mean
a
Adjusted
mean
a,b
P value
a,b
No lymphedema
464 192 (41) 48.6
Lymphedema 136 54 (40) 47.5
No lymphedema 464 186 (40) 48.5
Lymphedema, no burning pain
113 72 (64) 48.3 47.9
Lymphedema with burning pain 23 17 (74) 43.0
c
41.7 0.02
Abbreviation: MCS, mental component summary; MOS, medical outcomes study; SF, short form
a
Adjusted for age at 40-month assessment, race/ethnicity, disease stage, body mass index
b
Analysis of 136 lymphedema-affected women
c
Significantly different from women without lymphedema based on Scheffé’s pair-wise multiple comparison test
65
Table 9. The associations between self-reported lymphedema symptoms and perceived stress
Perceived Stress N Adjusted mean
a
Adjusted mean
a,b
P value
a,b
No lymphedema 464 8.9
Lymphedema 136 9.1
No lymphedema 464 8.9
Lymphedema, no burning pain 113 8.8 10.1
Lymphedema with burning pain 23 10.8
c,d
11.6 0.06
No lymphedema 464 8.9
Lymphedema, no tension 100 8.7 10.2
Lymphedema with tension 36 10.5
c,d
11.5 0.06
No lymphedema 464 8.9
Lymphedema, no raised temperature/redness 100 8.8 10.5
Lymphedema with raised temperature/redness 36 10.3
c,d
11.3 0.26
No lymphedema 464 8.9
Lymphedema, no dry skin 98 8.8
Lymphedema with dry skin 38 10.2
c
a
Adjusted for age at 40-month assessment, race/ethnicity, disease stage, body mass index
b
Mutually adjusted for the following symptoms: burning pain, tension and raised temperature/redness (analysis of 136 lymphedema-
affected women)
c
Significantly different from women without lymphedema based on Scheffé’s pair-wise multiple comparison test
d
Significantly different from lymphedema-affected women without a particular symptom based on Scheffé’s pair-wise multiple
comparison test
66
Table 10. Self-reported lymphedema symptoms by race/ethnicity and age at diagnosis
Self-reported lymphedema
symptom
Race/ethnicity Age at 40-month assessment (years)
Non-Hispanic white
N (%)
Black
N (%)
Hispanic white
N (%)
P value
a
38 to 49
N (%)
50 to 59
N (%)
60 to 69
N (%)
P value
a
No burning pain 55 (85) 43 (77) 15 (100) 36 (82) 49 (80) 28 (90)
Burning pain 10 (15) 13 (23) 0 (0) 0.09 8 (18) 12 (20) 3 (10) 0.46
No tension 51 (78) 35 (63) 14 (93) 30 (68) 46 (75) 24 (77)
Tension 14 (22) 21 (38) 1 (7) 0.03 14 (32) 15 (25) 7 (23) 0.61
a
P values were obtained from Chi-square tests
67
Chapter 4
Association of lymphedema with C-reactive protein and serum amyloid A
Breast cancer patients who have undergone treatment such as axillary lymph node
dissection or axillary radiotherapy tend to have compromised functioning of lymphatic
system. The lymphatic system plays a critical role in the maintenance of tissue fluid
homeostasis, immune responses, and absorption of dietary fats (1). Therefore, these
patients who have undergone axillary treatment are subject to accumulation of protein-
rich interstitial fluid and subcutaneous fat in the upper extremity (lymphedema) (2-4).
Despite the several measures to assess lymphedema, lymphedema is often
underdiagnosed or misdiagnosed (2,5).
Clinical assessment methods currently used for lymphedema include circumferential
measurement, water displacement, optoelectric volumetry, and bioimpedance
spectroscopy (6). Arm circumference, which is the most frequently used measurement,
can be measured with a tape measure. Although a tape measure is low-cost and portable,
the assessment can be time-consuming and lacks accuracy and reliability (7), especially
for early stage lymphedema (8,9). Other methods such as water displacement,
optoelectric volumetry, and bioimpedance spectroscopy may provide greater accuracy
than circumferential measurement, but these methods require special equipments that
might not be available at clinics. A low-cost, safe, and reliable assessment measure for
detecting and monitoring lymphedema is desirable.
68
C-reactive protein (CRP) and serum amyloid A (SAA), non-specific, acute-phase
proteins, have been used as a screening tool for inflammation, a marker for disease
activity, and a diagnostic adjunct for various infectious or inflammatory diseases (10,11).
Since lymphedema leads to subcutaneous accumulation of adipose tissue (5) which
produces inflammatory molecules (12,13), the levels of CRP and SAA concentrations
may be elevated in lymphedema-affected patients. To our knowledge, no studies have
examined the associations of CRP and SAA with lymphedema, and thus this preliminary
study explored the potential utility of these inflammatory biomarkers as a tool for
assessing breast cancer-related lymphedema by using data from a multi-ethnic cohort
study of female breast cancer survivors.
MATERIALS AND METHODS
Study setting, subjects, and recruitment
The aims, study design, and recruitment procedures of the HEAL Study have been
published previously (14,15). Briefly, the HEAL Study is a multi-center, multi-ethnic
prospective cohort study consisting of 1,183 female breast cancer survivors. Women
diagnosed with first primary in situ or Stage I-IIIA invasive breast cancer between 1995
and 1999 were recruited into the HEAL Study through the Surveillance, Epidemiology,
and End Results (SEER) registries in three regions of the United States: New Mexico,
Western Washington, and Los Angeles County, California. The age ranges studied varied
by study site with women 91 years of age or younger recruited in New Mexico, women
aged 40 to 65 years recruited in Western Washington, and women aged 35 to 64 years
recruited in Los Angeles County. A total of 1,183 women completed up to five
69
assessments over 10 years of follow up. Three of those assessments were used for this
study. The first assessment (baseline assessment) was administered in person within the
first year (on average, 6 months) after a woman’s diagnosis (Figure 3). The second
assessment was administered, on average, 30 months after a woman’s diagnosis (30-
month assessment) via in-person interview or self-completed questionnaire. The third
assessment was administered, on average, 40 months after a woman’s diagnosis (40-
month assessment) by telephone interview or mailed questionnaire in New Mexico, by
mailed questionnaire plus telephone follow-up in Western Washington, and by telephone
interview in Los Angeles County.
In the present study, we excluded 17 women who did not receive any surgery, leaving
1,166 women. A total of 933 (80%) of the 1,166 women completed the 30-month
assessment. Of these, 798 (86%) women provided a blood sample. A total of 716 (90%)
of the 798 women completed the 40-month assessment. We excluded 23 women with
incomplete data on weight or height (n=20) or date of lymphedema onset (n=3). The final
analytic cohort consisted of 693 women (Figure 4).
We obtained informed consent from all participants at each assessment. The study was
approved by the institutional review boards of participating centers, in accord with
assurances filed with and approved by the U.S. Department of Health and Human
Services.
Data collection
70
Measures of inflammation: CRP and SAA. A 30-ml 12-hour fasting blood sample was
collected at the 30-month assessment. All blood samples were processed and stored at -
70°C to -80°C until assays were conducted. High sensitivity CRP and SAA were
measured by latex-enhanced nephelometry using the Behring Nephelometer II analyzer
(Dade Behring Diagnostics, Deerfield, IL) at the University of Washington Medical
Center (Seattle, WA). The performance of this high sensitivity assay has been shown to
be good (16). The lower detection limit for CRP and SAA assays were 0.2 and 0.7 mg/l,
respectively. The inter-assay coefficients of variation were 5% to 9% for CRP and 4% to
8% for SAA. Quality control procedures were conducted by running each assay with
control materials from Bio-Rad Laboratories (Hercules, CA).
Lymphedema. Participants provided information on lymphedema at the 40-month
assessment. In the assessment, we described our definition of lymphedema to study
participants as follows: “Sometimes the arm on the side on which you had breast cancer
becomes swollen because of an accumulation of fluid in your arm. This is called
lymphedema. Please do not confuse this with the temporary swelling that occurs after
surgery.” We then asked the following “yes” or “no” question: “Have you experienced
lymphedema in your arm at any time since your breast cancer diagnosis?” Women who
answered “yes” also provided the month and year of their first onset of lymphedema.
Anthropometrics. Trained staff measured weight in a standard manner at the 30-month
assessment. Weight was measured to the nearest 0.1 kg using a balance-beam laboratory
scale at New Mexico and Western Washington, and a portable Thinner Digital Electronic
Scale at Los Angeles. All measurements were performed twice in succession. The
71
average of the two weight measurements was used for analyses. The baseline assessment
collected data on self-reported height at the age of 18 years. By using these data, we
calculated BMI as weight (kg) divided by the square of height (m
2
).
Lifestyle factors. The 30-month assessment captured information on a woman’s history of
smoking (never smoker, former smoker, active smoker) and sports and recreational
physical activity in the prior year. The level of sports and recreational physical activity
was measured by using the Modifiable Activity Questionnaire, which has demonstrated
acceptable validity and reliability(17). We calculated metabolic equivalent task (MET)
hours of physical activity based on the Compendium of Physical Activities compiled by
Ainsworth et al (18).
Other data. We obtained clinical data on age at diagnosis, disease stage (in situ,
localized, regional), and breast cancer treatment (chemotherapy, radiation therapy,
surgery type, and number of excised lymph nodes) from local SEER cancer registry
records and by abstracting participants’ hospital medical records. We used both self-
report and hospital medical records to determine tamoxifen use at the time of 30-month
assessment. History of comorbidities such as diabetes, myocardial infarction, heart
failure, kidney disease, arthritis, and hypertension was self-reported at the 30-month
assessment.
Statistical analysis
Distributions of demographic factors, breast cancer-related and treatment-related factors,
health-related factors, and lifestyle factors were compared between women who reported
72
lymphedema and women who did not by using Chi-square test or Fisher’s exact test. We
used analysis of variance (ANOVA) and analysis of covariance (ANCOVA) to compare
the levels of CRP and SAA by lymphedema status (present, absent). CRP and SAA
values were log transformed to better meet the assumptions of normality and linearity. By
using ANCOVA, we adjusted for race/ethnicity (Hispanic white or black, non-Hispanic
white or American Indian or Asian/Pacific Islander), and age (years) at the time of 30-
month assessment and BMI (kg/m
2
) at the time of 30-month assessment. We combined
Hispanic white women and black women because there was no significant difference in
concentrations of CRP and SAA between the two groups. We also combined the rest of
the racial/ethnic groups for the same reason. Additional adjustment of number of excised
lymph nodes or time since diagnosis did not alter the results substantially. We obtained P
values based on overall F statistics from ANOVA and ANCOVA, and also calculated
adjusted means of CRP and SAA for women with and without lymphedema.
RESULTS
A total of 120 (17%) of the 693 women reported lymphedema occurrence prior to 30-
month assessment (Table 11). Ten (8%) of the 120 women reported experiencing at least
one episode of infection associated with lymphedema (range: 1, 4). The median of CRP
was 2.2 mg/l (range: 0.2, 162) while the median of SAA was 9.6 mg/l (range: 0.7, 538).
These values were highly correlated with each other (correlation coefficient: 0.75,
P<0.001). The results indicated that breast cancer stage, type of surgery, chemotherapy,
number of excised lymph nodes, and BMI significantly differed by lymphedema status
(P<0.05) (Table 11).
73
The results indicated a statistically significant difference in the CRP level between
women with lymphedema and women without (P=0.03) (Table 12). The geometric means
of CRP level were 1.91 in women without lymphedema and 2.51 in women with
lymphedema. However, this difference was no longer significant after adjusting for race,
age, and BMI (P=0.52). No significant difference in the SAA level was observed between
women with lymphedema and women without before (P=0.60) or after controlling for
race, age, and BMI (P=1.00) (Table 12).
DISCUSSION
In this multi-ethnic cohort of 693 female breast cancer survivors who participated in the
HEAL Study, 120 (17%) women indicated that they experienced lymphedema prior to
the 30-month assessment. Our results indicated no significant associations between
lymphedema and the levels of CRP and SAA, suggesting that CRP and SAA may not be
useful in detecting or monitoring lymphedema in breast cancer survivors.
In our unadjusted analysis, women who indicated lymphedema had a significantly higher
concentration of CRP compared to women who did not report lymphedema; however, the
difference was no longer significant after adjusting for age, race/ethnicity, and BMI. In
our previous report, BMI greater than 30 kg/m
2
was a risk factor for developing
lymphedema. Moreover, another study which used data from the HEAL Study reported a
significant positive association between BMI and the concentration of CRP (13);
therefore, the significant difference we observed in our study is likely due to the
confounding effect of BMI.
74
One of the limitations of our study is that CRP and SAA were available only at one time
point in our study, thus we were unable to evaluate the changes over time. Studies with
repeated measures of CRP and SAA may be helpful in evaluating the utility of these
inflammatory markers for detecting or monitoring lymphedema. Furthermore, this study
solely relied on self-report to establish the presence of lymphedema and the self-report
was not verified using medical records. Although self-report was found to be highly both
sensitive and specific in a previous study (19), we cannot rule out the possibility of
misclassification.
In conclusion, the current study suggested that CRP and SAA measurements might not be
sensitive enough to identify breast cancer survivors with lymphedema. Given that
lymphedema can cause tremendous burden on breast cancer survivors, further
investigations of reliable methods to detect or monitor breast cancer-related lymphedema
are encouraged.
75
References
1. Norrmen C, Tammela T, Petrova TV, et al. Biological Basis of Therapeutic
Lymphangiogenesis. Circulation. 2011;123(12):1335–1351.
2. Rockson SG. Update on the Biology and Treatment of Lymphedema. Curr Treat
Options Cardio Med. 2012;14(2):184–192.
3. Mortimer PS. The pathophysiology of lymphedema. Cancer. 1998;83(12 Suppl
American):2798–2802.
4. Brorson H, Ohlin K, Olsson G, et al. Breast Cancer-Related Chronic Arm
Lymphedema Is Associated with Excess Adipose and Muscle Tissue. Lymphatic
Research and Biology. 2009;7(1):3–10.
5. Warren AG, Brorson H, Borud LJ, et al. Lymphedema. Annals of Plastic Surgery.
2007;59(4):464–472.
6. Smoot BJ, Wong JF, Dodd MJ. Comparison of diagnostic accuracy of clinical
measures of breast cancer-related lymphedema: area under the curve. Arch Phys
Med Rehabil. 2011;92(4):603–610.
7. Armer JM. The problem of post-breast cancer lymphedema: impact and
measurement issues. Cancer Invest. 2005;23(1):76–83.
8. Kosir MA, Rymal C, Koppolu P, et al. Surgical outcomes after breast cancer
surgery: measuring acute lymphedema. J. Surg. Res. 2001;95(2):147–151.
9. Armer JM, Radina ME, Porock D, et al. Predicting breast cancer-related
lymphedema using self-reported symptoms. Nurs Res. 2003;52(6):370–379.
10. Zhang N, Ahsan MH, Purchio AF, et al. Serum amyloid A-luciferase transgenic
mice: response to sepsis, acute arthritis, and contact hypersensitivity and the
effects of proteasome inhibition. 2005;174(12):8125–8134.
(http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=15944
321&retmode=ref&cmd=prlinks)
11. Clyne B, Olshaker JS. The C-reactive protein. J Emerg Med. 1999;17(6):1019–
1025.
12. Poitou C, Coussieu C, Rouault C, et al. Serum Amyloid A: A Marker of
Adiposity ‐ induced Low ‐ grade Inflammation but Not of Metabolic Status.
Obesity. 2006;14(2):309–318.
13. Pierce BL, Neuhouser ML, Wener MH, et al. Correlates of circulating C-reactive
protein and serum amyloid A concentrations in breast cancer survivors. Breast
Cancer Res Treat. 2009;114(1):155–167.
76
14. McTiernan A. Adiposity and Sex Hormones in Postmenopausal Breast Cancer
Survivors. Journal of Clinical Oncology. 2003;21(10):1961–1966.
15. Irwin ML, Crumley D, McTiernan A, et al. Physical activity levels before and after
a diagnosis of breast carcinoma. Cancer. 2003;97(7):1746–1757.
16. Ledue TB, Johnson AM, Cohen LA, et al. Evaluation of proficiency survey results
for serum immunoglobulins following the introduction of a new international
reference material for human serum proteins. Clin Chem. 1998;44(4):878–879.
17. Kriska AM, Caspersen CJ. Introduction to the collection of physical activity
questionnaires. In A Collection of Physical Activity Questionnaires for Health-
Related Research. Medicine & Science in Sports & Exercise. 1997;29(6
Suppl):S5–S9 and S73–S78.
18. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities:
an update of activity codes and MET intensities. Medicine & Science in Sports &
Exercise. 2000;32(9 Suppl):S498–504.
19. Norman SA, Localio AR, Potashnik SL, et al. Lymphedema in breast cancer
survivors: incidence, degree, time course, treatment, and symptoms. Journal of
Clinical Oncology. 2009;27(3):390–397.
77
Figure 4. Recruitment flow and data collection
Baseline assessment
(on average, 6 months following breast cancer
diagnosis)
- Sociodemographics: race/ethnicity
- Physical activity, smoking history
Hospital medical records (baseline data)/SEER
cancer registry records
- Breast cancer stage (SEER staging)
- Treatment (surgery, number of excised lymph
nodes, radiation, chemotherapy, tamoxifen)
Recruitment flow Data collection: Timing and variables
1,183 participants completed the baseline
assessment (original HEAL cohort).
- 615 New Mexico
- 202 Western Washington
- 366 Los Angeles County
Second assessment
(on average, 30 months following breast cancer
diagnosis)
- C-reactive protein, serum amyloid A
- Self-reported history of comorbidities
716 of the 798 participants who completed the
second assessment completed the third
assessment.
Exclusions, n (reason):
- 8 (deceased)
- 74 (did not complete the assessment
for other reasons)
Third assessment
(on average, 40 months following breast cancer
diagnosis)
- Self-reported lymphedema (yes, no)
- Date of first lymphedema onset (month, year)
- Number of episode of infection associated with
lymphedema
933 of the 1,166 participants who were eligible
for this analysis completed the second
assessment.
Exclusions, n (reason):
- 53 (deceased)
- 180 (did not complete the assessment
for other reasons)
798 of the 933 participants who completed the
second assessment also provided a blood sample.
693 of the 716 participants who completed the
third assessment comprised the analytic cohort.
Exclusions, n (reason):
- 20 (incomplete information of weight
or height)
- 3 (no date of lymphedema onset)
1,166 of the 1,183 participants who completed
the baseline assessment met the criteria for the
analysis.
Exclusions, n (reason):
- 17 (no surgery)
78
Table 11. Characteristics of 693 female breast cancer survivors by lymphedema
No lymphedema Lymphedema
N (%) N (%) P value
a
Number of participants 573 120
Age at 30-month assessment (years)
Mean (SD) 57.1 (9.9) 55.6 (10.0) 0.12
Race/ethnicity
Hispanic white 61 (80) 15 (20)
Non-Hispanic white 351 (84) 66 (16)
Black 140 (79) 38 (21)
Asian/Pacific Islander
American Indian, Other
21 (95) 1 (5) 0.14
Breast cancer stage
In situ 159 (96) 6 (4)
Localized 298 (79) 77 (21)
Regional 116 (76) 37 (24) <0.001
Surgery type
Partial/less than total
mastectomy/surgery, NOS
393 (85) 70 (15)
Total mastectomy/
modified radical mastectomy
180 (78) 50 (22) 0.03
Number of excised lymph nodes
c
Mean (SD) 9.0 (8.2) 14.1 (6.8) <0.001
Radiation therapy
No 235 (85) 41 (15)
Yes 338 (81) 79 (19) 0.16
Chemotherapy
No 408 (86) 65 (14)
Yes 165 (75) 55 (25) <0.001
Tamoxifen use at or before 30-month assessment
No 320 (84) 59 (16)
Yes 253 (81) 61 (19) 0.18
79
Table 11. Characteristics of 693 female breast cancer survivors by lymphedema (continued)
No lymphedema Lymphedema
N (%) N (%) P value
a
Body mass index at 30-month assessment (kg/m
2
)
Mean (SD) 27.3 (6.0) 29.3 (6.5) 0.001
History of comorbidities prior to 30-month assessment
No diabetes 512 (83) 108 (17)
Diabetes 61 (84) 12 (16) 0.83
No myocardial infarction 554 (83) 117 (17)
Myocardial infarction 19 (86) 3 (14) 0.22
b
No heart failure 562 (83) 116 (17)
Heart failure 11 (73) 4 (27) 0.31
b
No kidney disease 556 (83) 117 (17)
Kidney disease 17 (85) 3 (15) 1.00
b
No arthritis 345 (82) 77 (18)
Arthritis 228 (84) 43 (16) 0.42
No hypertension 375 (84) 74 (16)
Hypertension 198 (81) 46 (19) 0.43
Smoking status at 30-month assessment
Never, <0.5 packs of cigarettes in life 275 (83) 55 (17)
Former 230 (82) 49 (18)
Active smoker 68 (81) 16 (19) 0.87
Physical activity (MET hours/week)
<0.5 143 (84) 28 (16)
0.5–20.0 319 (84) 62 (16)
>20.0 111 (79) 30 (21) 0.38
Abbreviations: BMI, body mass index; MET, metabolic equivalent task; SD, standard deviation
a
P values were based on F test from ANOVA (continuous) or Chi-square test or Fisher’s exact test (categorical).
b
P values were Fisher’s exact test.
c
7 women who did not have information on number of excised lymph nodes were excluded.
80
Table 12. Geometric means of C-reactive protein (CRP) and serum amyloid A (SAA) by lymphedema status
Geometric mean CRP
P value
Geometric mean CRP
a
P value
a
Geometric mean
SAA
P value
Geometric mean
SAA
a
P value
a
No Lymphedema 1.91 1.97 5.87 5.87
Lymphedema 2.51 0.04 2.12 0.52 6.11 0.60 5.87 1.00
Abbreviation: CRP, C-reactive protein; SAA, serum amyloid A; SE, Standard error
a
The estimates were adjusted for race/ethnicity, age, and body mass index.
Abstract (if available)
Abstract
Lymphedema remains a significant issue among breast cancer survivors and thus, further efforts are needed to improve prevention and management of this complication. This dissertation explores lymphedema risk factors and the associations of lymphedema with quality of life (QOL) and inflammatory markers using data from a multi‐center, multi‐ethnic prospective cohort study of breast cancer survivors (Health, Eating, Activity and Lifestyle Study). A total of 1,183 women diagnosed with in situ or Stage I-IIIA breast cancer between 1995 and 1999 were recruited through the Surveillance, Epidemiology, and End Results registries in New Mexico, Los Angeles County, and Western Washington. Data were collected on sociodemographic factors, clinical characteristics, comorbidities, body mass index (BMI), hormonal factors, lifestyle factors, and QOL. A blood sample was collected to assess the levels of C‐reactive protein (CRP) and serum amyloid A (SAA) at about 30 months after diagnosis. The results indicated that extensive surgery/axillary lymph node dissection, obesity, chemotherapy, and hypertension increased the risk of lymphedema, and that lymphedema‐associated burning pain and tension decreased QOL. No significant associations between lymphedema and CRP/SAA were found after adjusting for BMI. These findings suggest the importance of monitoring survivors with lymphedema risk factors and relieving lymphedema‐associated burning pain and tension.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Genes and hormonal factors involved in the development or recurrence of breast cancer
PDF
DEXA measures of body fat percentage and acute phase proteins among breast cancer survivors
PDF
Predictive factors of breast cancer survival: a population-based study
PDF
Genetic risk factors in multiple myeloma
PDF
Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer
PDF
Association between informed decision-making and mental health-related quality of life in long term prostate cancer survivors
PDF
The role of heritability and genetic variation in cancer and cancer survival
PDF
Factors that influence mammographic density: role of estrogen metabolism genes, biomarkers of inflammation, and lifestyle
PDF
Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial
PDF
The effects of hormonal exposures on ovarian and breast cancer risk
PDF
Genetic and environmental risk factors for childhood cancer
PDF
Associations between exercise and quality of life in adult survivors of non-Hodgkin's lymphoma and colorectal cancer
PDF
An analysis of disease-free survival and overall survival in inflammatory breast cancer
PDF
Genetic correlates of the timing of natural menopause: the breast and prostate cancer cohort consortium
PDF
Screening and association testing of coding variation in steroid hormone coactivator and corepressor genes in relationship with breast cancer risk in multiple populations
PDF
Pathogenic variants in cancer predisposition genes and risk of non-breast multiple primary cancers in breast cancer patients
PDF
Body size and the risk of prostate cancer in the multiethnic cohort
PDF
Air pollution and breast cancer survival in California teachers: using address histories and individual-level data
PDF
The role of pesticide exposure in breast cancer
PDF
Genes and environment in prostate cancer risk and prognosis
Asset Metadata
Creator
Togawa, Kayo
(author)
Core Title
Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Epidemiology
Publication Date
04/29/2015
Defense Date
03/18/2014
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
arm morbidity,axillary lymph node dissection,body mass index,breast cancer survivors,breast cancer treatment,chemotherapy,C‐reactive protein,hypertension,inflammatory marker,lymphedema,OAI-PMH Harvest,Quality of life,risk factors,serum amyloid A
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Bernstein, Leslie (
committee chair
), Azen, Stanley P. (
committee member
), Mack, Wendy Jean (
committee member
), Mckean-Cowdin, Roberta (
committee member
), Winstein, Carolee J. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-407174
Unique identifier
UC11296494
Identifier
etd-TogawaKayo-2461.pdf (filename),usctheses-c3-407174 (legacy record id)
Legacy Identifier
etd-TogawaKayo-2461.pdf
Dmrecord
407174
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Togawa, Kayo
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
arm morbidity
axillary lymph node dissection
body mass index
breast cancer survivors
breast cancer treatment
chemotherapy
C‐reactive protein
hypertension
inflammatory marker
lymphedema
risk factors
serum amyloid A