136 Seminars in Oncology Nursing, Vol 23, No 2 (May), 2007: pp 136 –141

OBJECTIVES:

SYMPTOM CLUSTERS
To examine the current research
related to symptom clusters in
specific cancer diagnoses.

DATA SOURCES:
Research studies, review articles.
RELATED TO
CONCLUSION:
Factors that can impact symp-
SPECIFIC CANCERS
toms and their relationship in-
clude the stage of disease, treat-
ment used, modification of
treatment for the individual pa- AUDREY G. GIFT
tient, patient co-morbidities as
well as psychological, sociological,
and cultural factors. Understand-
ing is needed regarding which of
these factors individually or in
combination contribute to symp-
toms clustering together. The
mechanism underlying symptom
P atients interpret symptoms as an indication of the need
for health care. Health care professionals, on the other
hand, use symptoms as a first step in diagnosing a patient,
as a focus of treatment, and in assessing, monitoring, and/or
managing the disease process and its treatment. In advanced
perception and interpretation disease the focus is almost exclusively on the alleviation of symp-
need further exploration. Research toms. Recently there has been much written about symptom
focus should be on one type of clusters, but their usefulness in patient diagnosis, assessment,
cancer. monitoring, and management of care has not been adequately
examined. This article will address that issue as well as whether
IMPLICATIONS FOR NURSING clusters are useful as a predictor of future clinical events. Atten-
PRACTICE: tion will be on the research that has focused on a single type of
cancer, such as lung or breast cancer.
Symptom cluster research offers Most research related to symptom clusters has focused on pa-
little at this point to be helpful in
tients with cancer taken together as a group, rather than examin-
patient diagnosis, assessment,
ing the symptoms of a specific cancer, such as lung cancer or
monitoring, or management of
disease. breast cancer alone. For symptom clusters to be useful for prac-
tice, there needs to be more of a focus on symptom clusters in
different types of cancer. Those that have focused on one type of
KEYWORDS:
cancer identified symptoms normally seen in that type of cancer
Symptom cluster, lung cancer,
but may not have used the term “symptom cluster” or performed
breast cancer
the analysis to determine if the symptoms are related, a criterion
often used to define a symptom cluster.
Symptoms co-occur, especially in those with cancer, and they
may or may not be related to one another. At the time of diagnosis
the number of symptoms reported to co-occur vary from study to
Audrey G. Gift, PhD, RN, FAAN: Professor, study. Vainio, Auvinen, and members of the symptom prevalence
Michigan State University, College of Nurs- group,1 reported an average of only three symptoms at the time of
ing, East Lansing, MI.
diagnosis. In contrast, Gift et al2 found an average of 11 symptoms
Address correspondence to Audrey G.
Gift, PhD, RN, FAAN, Michigan State Uni- reported by patients with newly diagnosed lung cancer.
versity, College of Nursing, A212 Life Sci-
ences Bldg, East Lansing, MI 48824; SYMPTOMS BRING PATIENTS IN FOR HEALTH CARE
e-mail: [email protected]

© 2007 Elsevier Inc. All rights reserved.

0749-2081/07/2302-$30.00/0
P atients use symptoms as an indication that they are in need of
health care. Although not identified as a symptom cluster,
symptoms that constitute the warning signs of different types of
doi:10.1016/j.soncn.2007.01.003
cancer are well known. For breast cancer it is a swelling in a part
 SYMPTOM CLUSTERS RELATED TO SPECIFIC CANCERS 137

of the breast, skin irritation or dimpling, nipple derlying physical abnormality. The symptoms ex-
pain, or nipple turning inward, redness or scali- perienced by patients with lung cancer before
ness of the nipple or breast skin, nipple discharge, diagnosis did not discriminate between those with
or lump in the underarm area. But these symp- advanced disease and those with operable can-
toms have not been studied as a symptom cluster. cer.4 The cluster of symptoms experienced at the
Are the symptoms known to indicate the early time of cancer diagnosis are experienced in other
signs of a particular type of cancer a cluster? conditions and therefore do not serve to establish
There has been increasing attention given to a diagnosis. But they may serve as the catalyst for
teach the public the early signs of lung cancer, patients seeking health care.
such as a cough that does not go away and gets
worse over time, constant chest pain, coughing up
blood, shortness of breath, wheezing, hoarseness,
USING SYMPTOMS TO MONITOR DISEASE
repeated bouts of pneumonia or bronchitis, swell-
ing of the neck and face, loss of appetite or weight
loss, or fatigue. Yet patients do not always seek S ymptom clusters may not be helpful in moni-
toring symptoms in specific cancers. Sarna
and Brecht5 explored a combination of symptoms
health care even when they experience new symp-
toms. Most will manage the symptoms themselves in 60 women with advanced lung cancer and
for a long time before entering the health care found fatigue, disruptions in outlook, frequent
system, even though they may be at high risk for pain, and difficulties in sleeping were rated the
cancer. most distressing and were the most prevalent se-
rious symptom disruptions. Five factors, or clus-
ters, were found and labeled as emotional suffer-
SYMPTOMS IN DIAGNOSIS ing, physical suffering, gastrointestinal distress,
respiratory distress, and malaise. In another study

T he most common use of symptom clusters is

for diagnosis. An example is a study by Ham-
ilton et al3 in which they compared 247 primary
of 69 women with primary or recurrent lung can-
cer, Sarna6 found the most prevalent and most
distressing symptoms to include fatigue, frequent
lung cancer patients in 21 general practices with pain, and insomnia, not symptoms specific to lung
1,235 controls matched on age, sex, and general cancer.
practice and found seven symptoms (hemoptysis, Gift et al,2 rather than examining the symptoms
loss of weight, loss of appetite, dyspnea, thoracic that were the most distressing, used factor analy-
pain, fatigue, and cough) to be associated with sis on symptom severity to identify a cluster of
later development of lung cancer. Thus, this clus- symptoms in patients newly diagnosed with lung
ter of symptoms will identify those patients likely cancer. The symptoms that fell into the cluster
to develop lung cancer in the future. were fatigue, weakness, nausea, vomiting, loss of
A research team that explored the early symp- appetite, loss of weight, and altered taste. In both
toms of lung cancer asked newly diagnosed lung the Gift and Sarna studies, the symptoms identi-
cancer patients to recall the symptoms experi- fied as part of the cluster are not those specific to
enced before diagnosis and compared them with lung cancer, but rather could be reported by pa-
patients awaiting surgery. Most reported having tients with any type of cancer.
experienced new symptoms such as cough, Those who have focused on the symptoms in
breathing changes, and pain in the chest in the breast cancer have looked at the relationship
year before diagnosis. Other new symptoms re- among the symptoms experienced but have not
ported were fatigue, lethargy, weight loss, and explored the notion of symptom clusters. Gatson-
eating changes. Patients reported attempting to Johansson et al7 studied patients with breast can-
manage their symptoms on their own for as long cer and reported that fatigue, pain, and depression
as 2 years before seeking health care. The pa- significantly related to one another and to total
tients interpreted their symptoms as signs of health status. Jacobsen et al8 studied 54 women
aging rather than disease (with the exception of with breast cancer and reported that symptoms
hemoptysis), even though they were smokers, such as pain, emotional upset, nausea, mouth
making them at high risk for lung cancer.4 sores, muscle weakness, change in taste, sleep
Symptom severity ⬎or the number of symp- problems, chills, and vomiting were associated
toms reported do not always match with the un- with significant increases in fatigue severity. Bro-
138 A.G. GIFT

eckel et al,9 in their study of 61 women with breast pressed or blue, feeling anxious or nervous, and
cancer, found that these patients experienced loss of concentration. Thus, even in symptom
more severe fatigue symptoms that were signifi- cluster research the symptoms found in the clus-
cantly related to poorer sleep quality, more meno- ter are not those specific to the type of cancer
pausal symptoms, greater use of catastrophizing studied, but rather those that can occur in any
as a coping strategy, and current presence of a type of cancer.
psychiatric disorder in comparison with women
with no history of cancer. In a review of the IMPLICATIONS FOR ASSESSMENT OF PATIENTS
literature, Wilmoth et al10 note that patients with
breast cancer have been found to experience fa-
tigue, weight gain, and altered sexuality. They
note that these symptoms have not been explored
T he first step in symptom monitoring is to
identify an appropriate measure. Because
symptoms are subjective experiences, the only
as a cluster.
feasible indicator of symptoms is patient self-re-
Bower et al,11 in a study of 1,957 women with
port. To focus on symptom clusters it is essential
breast cancer, likewise found fatigue not only to
to have a measure that includes multiple symp-
be related to pain and sleep disturbance but also
toms. However, most of the measures that include
closely linked to depression. They did not find
multiple symptoms, such as the Memorial Symp-
fatigue to be related to menopausal symptoms.
tom Assessment Scale,16 are lengthy and not prac-
Thus, the symptoms found to be related to one
tical for clinical use. What is needed is a measure
another in patients with breast cancer, such as
that includes only the most important symptoms.
fatigue, pain, sleep disturbance, and depression,
But research has not yet identified what those
are found in a wide variety of cancers, not just
symptoms are.
breast cancer.
The descriptors used to report a symptom have
Part of using symptoms to monitor disease is
also been found to connote different sensations
learning how symptoms change over the course of
that are important for a differential diagnosis.
disease. A study of the lung cancer disease pro-
When different words are used it may result in
cess12 found symptoms to be greater in those with
different sensations being measured, making com-
more advanced disease. Another study13 focused
parisons across studies difficult. Also, grouping
on symptom clusters in 117 patients with lung
symptoms into physical symptoms (eg, pain, nau-
cancer over the course of the disease, and re-
sea, diarrhea, fatigue, wasting/cachexia), cognitive
ported that symptom distress was moderately high
symptoms (eg, poor problem solving, memory,
at the time of diagnosis, decreased 3 months after
attention), and affective symptoms (eg, anxiety
diagnosis, and then increased at 6 months. Those
and depression) has not proven helpful in under-
with more disease reported a higher number of
standing symptom clusters.
symptoms. But, Gift et al,14 in their study of a
symptom cluster (fatigue, weakness, nausea, vom-
iting, loss of appetite, loss of weight, and altered IMPLICATIONS FOR MANAGEMENT
taste), in patients with lung cancer at intervals for
a year following diagnosis found symptom severity
to decrease over the year.
Bender et al15 conducted a secondary analysis
I t is important to treat symptoms as a group, not
simply focusing on one symptom. Sarna17 found
the use of a structured nursing tool assessing
of three independent studies to explore symptom symptom distress to be effective in lowering symp-
clusters across three phases of breast cancer. The tom distress over time. Evidently, having the
phases included post-primary surgery and before nurse assess symptoms prompts them to inter-
adjuvant therapy in woman with early stage breast vene in an effective manner.
cancer, post-adjuvant therapy in women with Recently Solà et al18 did a Cochrane Review to
early stage breast cancer, and palliative care of assess the effectiveness of non-invasive interven-
women with metastatic breast cancer. They iden- tions delivered by health care professionals in
tified three symptom clusters corresponding to improving physical and psychological symptoms
three phases of breast cancer experience. Six in patients with lung cancer. They noted that two
symptoms that were common across all three trials of a nursing intervention resulted in im-
clusters included fatigue, lack of energy, de- proved breathlessness, performance status, and
creased physical strength/weakness, feeling de- emotional functioning.19,20 Three trials assessed
 SYMPTOM CLUSTERS RELATED TO SPECIFIC CANCERS 139

structured nursing programs and found positive precipitate certain symptoms, provide the setting
effects on delay in clinical deterioration, depen- for symptoms to interact with each other, and
dency and symptoms distress, and improvements confound our understanding of symptom clusters.
in emotional functioning and satisfaction with Without teasing apart these multiple co-occurring
care. One trial of reflexology showed some posi- contributors to the symptom experience, our un-
tive, but short-lasting effects on anxiety. derstanding of symptom clusters will remain in-
Psychological symptoms also require manage- complete.
ment. Mortazeri et al21 noted the need for psycho- Little is known about the relationships among
social interventions in those patients with lung symptoms. When symptoms co-occur they could
cancer and depression. They found that depres- escalate with the addition of each symptom,24
sion was predicted by having high symptom sever- they could be interactive,25 or the presence of a
ity, low social functioning, and being the recipient particular symptom, such as fatigue, could make
of palliative treatment. Linn et al22 demonstrated all other symptoms seem worse. However, without
the effectiveness of individual counseling sessions this key symptom, the symptom experience would
in decreasing depressive symptoms. be perceived as less severe. Duration and interpre-
Treatment for one symptom may “cross-over” tation of symptoms are other factors that influence a
and reduce the burden of other symptoms. Fleish- patient’s symptom reporting. Few researchers have
man23 noted the importance of treating pain, de- studied these types of relationships in the context of
pressive symptoms, and fatigue in tandem with a specific diagnosis.
efforts to cure or contain the cancer. Therapies Researchers have not included social or cogni-
include nutrition, psychosocial and cognitive-be- tive factors in their model, yet it is well known
havior therapies, education, coping strategies, and that patients with cancer are vigilant in their
exercise, as well as conventional pharmacologic monitoring of their symptoms, especially after
interventions. They make a case for treating treatment. Cancer is not considered to be cured
symptoms as a cluster, rather than independently, until after the 5-year marker. Patients monitor
and support the usefulness of symptom clusters in their symptoms and may interpret a symptom
the management of patients with cancer. such as a cough very differently after a diagnosis of
There is much in the literature about the symp- cancer than they would have before the diagnosis.
toms experienced at different times during the They have a tendency to report more symptoms
course of the disease (such as at diagnosis, after because of their fear that every little symptom
treatment, or with advanced disease), as well as may be an indication of returning disease. Thus,
the symptoms in different types of cancer (lung, symptom management after treatment needs to
breast, or colon cancer), during different treat- be handled differently than during treatment.
ments (such as chemotherapy, radiation, and sur- Another way to explore the relationship among
gery). A study of 60 women with lung cancer symptoms is to examine their trajectory over
emphasized the importance of knowing the ther- time. It would be important to note if different
apeutic regimen of the patient because patients symptoms have different trajectories over time.
receiving radiation therapy had different symptoms Stommel et al26 observed depression scores over
compared with those receiving chemotherapy.5 But time in patients with breast, colon, lung, and
it is even more complicated because often treat- prostate cancer, and found that those with lung
ments such as chemotherapy are modified or mod- cancer had higher depression scores than those
ulated according to the tolerance of the individual with other cancers and higher scores on other
patient creating much variability even within a che- somatic symptoms as well. Looking at symptoms
motherapeutic protocol. over time, they found that the depression and
In addition to the cancer and its treatment, somatic symptom scores declined over time while
another consideration is the co-morbidities and the patient’s sense of well being did not, demon-
personal characteristics of the patients. Sarna6 strating that depression and well being are not
recommends assessing for concurrent respiratory opposite ends of the same continuum, but rather
disease, previous chemotherapy, recurrent lung are independent of each other and thus not a
cancer, lack of surgical interventions, and low cluster.
income because they were found to be associated Knowledge of the etiology and mechanisms for
with high levels of symptom distress in women symptoms is an area for future research to ad-
with lung cancer. It is these various factors that vance our knowledge of symptom clusters. Symp-
140 A.G. GIFT

toms seen in cancer have been likened to the ter, but there is evidence to support the notion of
“sickness syndrome,”27 which proposes a com- symptom distress as a separate important symp-
mon etiology for the symptoms. Also, there are tom dimension.29 Different patients find different
similar areas of the brain in which symptoms such symptoms to be highly distressing, even though
as pain and dyspnea are interpreted. Further ex- they may not be the most severe. There may be
ploration of symptom mechanisms needs to be other dimensions of symptoms, such as duration,
combined with our exploration of symptom clus- that may be more meaningful in the formation of
ters. a cluster. Also, different symptom measures use
different words to describe a symptom, such as
CONCLUSION using “breathing difficulty” rather than “shortness
of breath,” and there is research to show that
different words can mean different sensations to
T he study of symptoms and their relationship
to one another forming symptom clusters is
an interesting area of study. At this point, how-
patients.30 This needs to be observed when exam-
ining the symptom cluster research.
ever, its usefulness in patient diagnosis, assess- There are psychological influences on the
ment, monitoring, or management of specific di- symptoms experienced. Having had a prior bout
agnoses is not well understood. This is largely with cancer could make the person hyper-vigilant
because of factors in the patient setting that im- in monitoring their symptoms, especially those
pact symptoms, most importantly the stage of thought to indicate a return of the cancer. Having
disease, treatment used, how that treatment is a family history or personal experience ridden
modified for the individual patient, and the pa- with high symptom intensity related to cancer
tients’ other co-morbidities. Thus, we do not know could alter one’s perception of symptoms as well.
which of these factors individually or in combina- Cultural factors may also influence how symp-
tion contribute to symptoms forming a cluster. toms are reported and how they relate to one
Research is needed that can tease apart these another.
factors to study their influence separately and Symptom cluster research thus far has paid
then in relation to one another. little attention to the mechanism of the symp-
The clusters that have been identified tend to toms. This is an important area for future research
contain symptoms that are general in nature, to aid our understanding of symptom clusters.
rather than those specific to the type of cancer Perhaps when a common area of the brain is found
involved. This is especially surprising in those to receive the signal from two symptoms they are
studies that include patients with only one type of more likely to cluster together. Also, the influence
cancer. of cytokines in the formation of a symptom cluster
There are also issues in the measurement of needs further study. It is important to explore
symptom clusters.28 Often intensity or severity is these mechanisms to learn more about how and
the only dimension considered in forming a clus- why symptoms cluster together.

REFERENCES

1. Vainio A, Auvinen A, with Members of the Symptom 5. Sarna L, Brecht ML. Dimensions of symptom distress in
Prevalence Group. Prevalence of symptoms among patients women with advanced lung cancer: a factor analysis. Heart
with advanced cancer: an international collaborative study. Lung 1997;26:23-30.
J Pain Symptom Manage 1996;12:3-10. 6. Sarna L. Correlates of symptom distress in women with
2. Gift AG, Jablonski A, Stommel M, et al. Symptom clusters lung cancer. Cancer Pract 1993;1:21-28.
in elderly patients with lung cancer. Oncol Nurs Forum 7. Gatson-Johansson F, Fall-Dickson JM, Bakos AB, et al.
2004;31:203-209. Fatigue, pain and depression pre-auto-transplant breast can-
3. Hamilton W, Peters TJ, Round A, et al. What are cer. Cancer Pract 1999;7:240-247.
the clinical features of lung cancer before the diagnosis is 8. Jacobsen PB, Hann DM, Azzarello LM, et al. Fatigue in
made? A population based case-control study. Thorax women receiving adjuvant chemotherapy for breast cancer:
2005;60:1059-1065. characteristics, course and correlates. J Pain Symptom Manage
4. Corner J, Hopkinson J, Fitzsimmons D, et al. Is late 1999;18:233-242.
diagnosis of lung cancer inevitable? Interview study of pa- 9. Broeckel JA, Jacobsen PB, Horton J, et al. Characteristics
tients’ recollections of symptoms before diagnosis. Thorax and correlates of fatigue after adjuvant chemotherapy for
2005;60:314-319. breast cancer. J Clin Oncol 1998;16:1689-1696.
 SYMPTOM CLUSTERS RELATED TO SPECIFIC CANCERS 141

10. Wilmoth MC, Coleman EA, Smith SC, et al. Fatigue, 20. Moore S, Corner J, Haviland J, et al. Nurse led follow
weight gain and altered sexuality in patients with breast can- up and conventional medical follow up of patients with lung
cer: exploration of a symptom cluster. Oncol Nurs Forum cancer: randomized trial. BMJ 2002;325:1145-1151.
2004;31:1069-1073. 21. Mortazeri A, Milroy R, Hole D, et al. Quality of life in
11. Bower JE, Ganz PA, Desmond KA, et al. Fatigue in breast lung cancer patients: as an important prognostic factor. Lung
cancer survivors: occurrence, correlates, and impact on qual- Cancer 2001;31:233-240.
ity of life. J Clin Oncol 2000;18:743-753.
12. Hopewood P, Stevens RJ. Symptoms at presentation for 22. Linn M, Linn B, Harris R. Effects of counseling for late
treatment in patients with lung cancer: implications for the stage cancer patients. Cancer 1982;16:1048-1055.
evaluation of palliative treatment. Br J Cancer 1995;71: 23. Fleishman SB. Treatment of symptom clusters: pain,
633-636. depression, and fatigue. J Natl Cancer Inst Monogr 2004;
13. Cooley ME, Short TH, Moriarty HJ. Patterns of symptom 32:119-123.
distress in adults receiving treatment for lung cancer. J Palliat 24. Lenz E, Pugh L, Milligan R, et al. The middle-range
Care 2002;18:150-159. theory of unpleasant symptoms: an update. Adv Nurs Sci
14. Gift AG, Stommel M, Jablonski A, et al. A cluster of 1997;19:14-27.
symptoms over time in patients with lung cancer. Nurs Res
25. Parker KP, Kimble LP, Dunbar SB, et al. Symptom
2003;52:393-400.
interactions as mechanisms underlying symptom pairs and
15. Bender CM, Ergun FS, Rodenzweig MQ, et al. Symptom
clusters. J Nurs Scholar 2005;37:209-215.
clusters in breast cancer across 3 phases of the disease. Cancer
Nurs 2005;28:219-225. 26. Stommel M, Kurtz ME, Kurtz JC, et al. The longitudinal
16. Portenoy RK, Thaler HT, Kornblith AB, et al. The Me- analysis of the course of depressive symptomatology in geriat-
morial Symptom Assessment Scale: an instrument for the ric patients with cancer of the breast, colon, lung, or prostate.
evaluation of symptom prevalence, characteristics and dis- Health Psychol 2004;23:564-573.
tress. Eur J Cancer 1994;30A:1326-1336. 27. Cleeland CS, Bannett GJ, Dantzer R, et al. Are the
17. Sarna L. Effectiveness of structured nursing assessment symptoms of cancer and cancer treatment due to a shared
of symptom distress in advanced lung cancer. Oncol Nurs biologic mechanism? Cancer 2003;97:2919-2925.
Forum 1998;25:1041-1048.
28. Barsevick AM, Whitmer K, Nail LM, et al. Symptom
18. Solà I, Thompson E, Subirana M, et al. Non-invasive inter-
cluster research: conceptual, design, measurement, and anal-
ventions for improving well-being and quality of life in patients
ysis issues. J Pain Symptom Manage 2005;31:85-95.
with lung cancer. The Cochrane Database of Systematic Reviews.
Chichester, UK: John Wiley & Sons: 2006. Available at: http:// 29. Tishelman C, Degner LF, Rudman A, et al. Symptoms in
www.mrw.interscience.wiley.com/cochrane/clsysrev/articles/ patients with lung carcinoma: distinguishing distress from in-
CD004282/frame.html. (accessed May 19, 2006). tensity. Cancer 104;9:2013-2021.
19. Corner J, Plant H, A’Hern R, et al. Non-pharmacological 30. Harver A, Mahler DA, Schwartzstein RM, et al. Descrip-
intervention for breathlessness in lung cancer. Palliat Med tors of breathlessness in healthy individuals: distinct and sep-
1996;10:299-305. arable constructs. Chest 2000;118:679-690.