Breast Cancer Res Treat (2012) 134:291–298

DOI 10.1007/s10549-012-2039-z

CLINICAL TRIAL

High-sensitivity C-reactive protein (hs-CRP) as a biomarker

for trastuzumab-induced cardiotoxicity in HER2-positive
early-stage breast cancer: a pilot study
Adedayo A. Onitilo • Jessica M. Engel •

Rachel V. Stankowski • Hong Liang •

Richard L. Berg • Suhail A. R. Doi

Received: 9 February 2012 / Accepted: 15 March 2012 / Published online: 4 April 2012
Ó Springer Science+Business Media, LLC. 2012

Abstract Monitoring of left ventricular ejection fraction intervals during trastuzumab treatment. Laboratory testing for
(LVEF) is the current standard for detection of trastuzumab- candidate biomarkers was repeated every 3 weeks with each
induced cardiotoxicity; however, time-to-diagnosis and cost cycle of trastuzumab. Trastuzumab-induced cardiotoxicity
of assessment are suboptimal in women with early-stage was defined as a decrease in LVEF of C15 % or to a value
breast cancer. We assessed the utility of B-type natriuretic below 50 %. A clinically significant decrease in LVEF was
peptide (BNP), high-sensitivity C-reactive protein (hs-CRP), observed in 28.6 % of women. Abnormal hs-CRP (C3 mg/L)
and cardiac troponin I (cTnI) as serum biomarkers for early predicted decreased LVEF with a sensitivity of 92.9 % (95 %
detection of trastuzumab-induced cardiotoxicity. Fifty-four CI 66.1–99.8) and specificity of 45.7 % (95 % CI 28.8–63.4),
women with human epidermal growth factor receptor 2 and subjects with normal hs-CRP levels (\3 mg/L) have
(HER2)-positive early-stage breast cancer were prospectively 94.1 % negative predictive 94.1 % (95 % CI 70.3–99.9)
enrolled, and the relationship between elevated serum BNP, suggesting that normal hs-CRP levels may be associated with
hs-CRP, and cTnI levels and clinically significant decreases in low future risk for decreased LVEF; however, no association
LVEF was examined. LVEF was monitored at 3–4 month with BNP or cTnI was observed. A false positive would have a
relatively low associated cost in breast cancer patients
undergoing adjuvant trastuzumab therapy and would indicate
A. A. Onitilo (&)
Department of Hematology/Oncology, Marshfield Clinic Weston continuation of routine observation during treatment through
Center, 3501 Cranberry Boulevard, Weston, WI 54476, USA traditional means. The maximum hs-CRP value was observed
e-mail: [email protected] a median of 78 days prior to detection of cardiotoxicity by
decreased LVEF, and those with normal levels were at lower
A. A. Onitilo
R. V. Stankowski
Marshfield Clinic Research Foundation, Marshfield, WI, USA risk for cardiotoxicity. Regular monitoring of hs-CRP holds
e-mail: [email protected] promise as a biomarker for identifying women with early-
stage breast cancer at low risk for asymptomatic trastuzumab-
A. A. Onitilo
S. A. R. Doi
induced cardiotoxicity. To our knowledge, this is the first
Clinical Epidemiology Unit, School of Population Health,
University of Queensland, Brisbane, QLD, Australia study documenting the utility of a less expensive, reproduc-
e-mail: [email protected] ible, easily obtainable biomarker with rapid results for eval-
uating cardiotoxicity related to trastuzumab therapy.
J. M. Engel
Department of Hematology/Oncology, Marshfield Clinic Cancer
Care at St Michaels, Stevens Point, WI, USA Keywords Antibodies, Monoclonal/therapeutic use

e-mail: [email protected] Breast neoplasms/metabolism/mortality
Heart diseases/
chemically induced
hs-CRP
Receptor, HER2
H. Liang
R. L. Berg
Biomedical Informatics Research Center, Marshfield Clinic
Research Foundation, Marshfield, WI, USA Introduction
e-mail: [email protected]
R. L. Berg Incorporation of trastuzumab (HerceptinÒ; Roche) into
e-mail: [email protected] treatment regimens represents a major advancement in the

123
292 Breast Cancer Res Treat (2012) 134:291–298

treatment of human epidermal growth factor receptor 2 accurate, simple, and more cost-effective way of assessing
(HER2)-positive breast tumors by specifically targeting the early cardiac dysfunction. The identification of markers of
HER2 pathway [1]. HER2-positive tumors are associated early cardiotoxicity is especially important due to the
with a more aggressive phenotype and poorer prognosis reversibility of early or asymptomatic trastuzumab-induced
[2, 3]. The monoclonal antibody trastuzumab is an adjuvant cardiotoxicity [3, 8].
agent that binds to the extracellular domain of the HER2 In this study, the utility of B-type natriuretic peptide
protein, regulating cell growth, differentiation, and sur- (BNP), high-sensitivity C-reactive protein (hs-CRP), and
vival. A meta-analysis of the pivotal randomized controlled cardiac troponin I (cTnI) as potential biomarkers for the
trials of adjuvant trastuzumab for treatment of women with detection of trastuzumab-induced cardiotoxicity in early-
HER2-positive metastatic breast cancer demonstrated sig- stage breast cancer patients was assessed. BNP is synthe-
nificant reduction of mortality, recurrence, and metastases sized in the ventricles of the heart and secreted in response
rates [4], and longer follow-up data has shown improve- to volume expansion and pressure overload; CRP is an
ment in disease-free survival [3]. acute phase protein produced by hepatocytes during peri-
Treatment with trastuzumab is generally well-tolerated. ods of inflammation; and cTnI is located in the contractile
The risk for trastuzumab-induced cardiotoxicity is the pri- apparatus of myocytes and is released upon necrosis.
mary morbidity of concern. Cells of the myocardium have Several studies have demonstrated that BNP, hs-CRP, and/
limited regenerative capacity, and damage can be perma- or cTnI elevation correlates with cardiac dysfunction
nent. The extent of cardiomyopathy secondary to trast- ranging in severity from asymptomatic decrease in LVEF
uzumab treatment ranges from an asymptomatic decrease in to heat failure [10–17]. We hypothesized that elevated
left ventricular ejection fraction (LVEF) to heart failure [5]. levels of serum BNP, hs-CRP, and/or cTnI may be indic-
The risk for trastuzumab-induced cardiotoxicity increases ative of trastuzumab-induced cardiotoxicity. Herein, we
with concurrent chemotherapy and increasing age, and report that hs-CRP shows promise as a biomarker for early
additive risk is incurred by anthracycline exposure, chest identification of patients experiencing asymptomatic trast-
wall irradiation, and pre-existing cardiac dysfunction [3, 6]. uzumab-induced cardiotoxicity.
Trastuzumab-induced cardiotoxicity does not appear to be
dose- or duration-dependent [7] and may be reversed by
withholding or discontinuing trastuzumab treatment with Materials and methods
the support of standard medical cardiac therapy [8].
Asymptomatic declines in LVEF constitute the majority Patient selection and data collection
of cardiac events observed with trastuzumab therapy [5]. In
a review of outcomes from several phase II and III che- Female patients ages 18 years and older were prospectively
motherapy trials, Seidman et al. [9] reported that the inci- enrolled at the Marshfield Clinic Cancer Care System begin-
dence of cardiomyopathy following treatment with ning in March 2007. Approval was granted by the Marshfield
trastuzumab alone ranged from 3–7 %, and was as high as Clinic institutional review board, and patients were enrolled
27 % when trastuzumab was administered in conjunction following first diagnosis of HER2-positive primary breast
with anthracyclines. Similarly, upon meta-analysis of the cancer (stages 1–3) if deemed eligible for trastuzumab adju-
pivotal randomized controlled trials, Viani et al. [4] found vant treatment. After informed consent was obtained from the
that the risk for cardiotoxicity was increased by 2.45-fold participant, baseline levels of serum BNP, hs-CRP, and cTnI
(95 % CI 1.89–3.16) in trastuzumab study arms. Therefore, were determined, and measurements were repeated every
periodic cardiac evaluation and monitoring is routine during 3 weeks during subsequent treatment up to 1 year. Patients
trastuzumab treatment. Cardiac monitoring and therapeutic were not excluded based on elevated levels of serum BNP, hs-
decisions are currently based on LVEF results from echo- CRP, or cTnI at the time of enrollment. Fifty-four subjects had
cardiogram (ECHO) or multiple gated acquisition (MUGA) been enrolled in the cohort as of February 2011.
radionuclide scanning performed every 3–4 months post- BNP was measured using the TriageÒ BNP Test (Inver-
treatment initiation. While these tests provide important ness Medical, Princeton, NJ), with concentrations reported to
information, they may not always detect early myocardial the nearest whole number in pg/mL. For study purposes,
damage. Even in the most proactive surveillance plans, values C200 pg/mL were considered abnormal [11]. Levels
these cardiac monitoring tests are done at intervals after of hs-CRP were measured using latex-enhanced nephelom-
several treatments of trastuzumab have been administered etry, reported in mg/L to the nearest tenth in the range of
when serious damage may have already occurred. Addi- 0.2–100.0 mg/L. For study purposes, values C3 mg/L were
tionally, serial assessment of cardiac function with ECHO/ considered abnormal [18]. Levels of cTnI were measured
MUGA after chemotherapy is elaborate and expensive. using the Dimension Vista assay (Siemens Healthcare
These limitations create the need for a rapid, reliable, Diagnostics, Deerfield, IL) with concentrations reported to

123
Breast Cancer Res Treat (2012) 134:291–298 293

the nearest tenth in ng/mL. For study purposes values of treatment were enrolled in this study. Five patients were
C0.1 ng/mL were considered abnormal [19]. excluded from analysis due to missing laboratory values or
Standard-of-care serial ECHO or MUGA scans were insufficient testing for LVEF (Fig. 1). Patients were divi-
performed to assess left ventricular contractile function ded into two groups based on observed cardiotoxicity, as
every 3–4 months during treatment, per the treating defined above in terms of declining LVEF.
oncologist. Baseline ECHO or MUGA scans were ideally In this study, 28.6 % (14/49) of women treated with trast-
performed in the time frame 1 month pre- to 1 month post- uzumab experienced a clinically important decrease in LVEF
trastuzumab initiation. For study purposes, clinically following initiation of treatment. There were no reports of
important cardiotoxicity was defined as a decrease in symptomatic heart failure. Trastuzumab treatment was dis-
ejection fraction (EF) of 15 % or more from baseline or to a continued in one subject and briefly held and then restarted in
value below 50 %. All other patients were categorized as 4 subjects following clinically significant decrease in LVEF.
normal LVEF. Clinically relevant changes were acted upon Subjects ranged in age from 30 to 90 years in the normal
by best clinical practice per the managing oncologist. LVEF group and 42 to 83 years in the decreased LVEF group.
Physicians reading ECHO/MUGA results were blinded to There was no significant difference between those with nor-
laboratory results for BNP, hs-CRP, and cTnI. Associations mal and decreased LVEF in age or BMI (Table 1) or fre-
between cardiotoxicity and abnormal laboratory results quency of other known risk factors for cardiac dysfunction
were evaluated using only those laboratory results collected including neoadjuvant chemotherapy, radiation therapy,
before detection of cardiotoxicity by change in LVEF. treatment with anthracycline, taxane, or endocrine therapies,
hypertension, diabetes, hypercholesterolemia, CAD, smok-
Statistical analysis ing, use of HRT/OCT, or use of acetylsalicylic acid, statins, or
antihypertensives (Table 2). Patients with decreased LVEF
The primary aim of the data analysis was to evaluate the had fewer ECHO/MUGA and laboratory tests as a result of the
relationship between BNP, hs-CRP, and cTnI levels and development of cardiotoxicity (Table 1).
LVEF as assessed by ECHO/MUGA. Wilcoxon’s rank- Plasma cTnI levels were consistently below the upper
sum tests were used to assess the differences in continuous limit of normal (\0.1 ng/ml), showing no detectable change
variables between the groups with normal and decreased over time, and were removed from further analysis. Patients
LVEF. The Pearson’s Chi-square test and Fisher’s exact with decreased LVEF had significantly elevated maximum
test were performed to evaluate the differences in discrete hs-CRP values pre-toxicity (before decrease in LVEF)
variables. A secondary multivariable logistic regression compared to patients with normal LVEF (Fig. 2a), but there
analysis using a forward selection method was conducted were no significant differences in maximum BNP values
to evaluate potential risk factors for decreased LVEF pre-toxicity (Fig. 2b). Only maximum hs-CRP was pre-
including maximum pre-toxicity hs-CRP, any abnormal hs- dictive of cardiotoxicity, not maximum BNP (Fig. 3). The
CRP (C3 mg/L) pre-toxicity, age, body mass index (BMI), three most predictive parameters for our outcome in the
neoadjuvant chemotherapy, radiation therapy, anthracy- final multivariable logistic regression model were an
cline or taxane use, endocrine therapy, hypertension, dia- abnormal hs-CRP value (C3 mg/L) pre-toxicity, presence
betes, hypercholesterolemia, coronary artery disease of CAD, and use of anthracycline chemotherapy (trend), but
(CAD), smoking, hormone replacement therapy/oral con- not abnormal BNP (C200 pg/mL) (Table 3). When abnor-
traceptive pills (HRT/OCP), or use of acetylsalicylic acid, mal hs-CRP and BNP values were considered together,
statins, or antihypertensives. Due to the small sample size, there was no significant improvement in test characteristics.
a low threshold was set for inclusion of variables in the
final model at a significance level of 0.2. Area under the
curve (AUC) was used to evaluate the clinical performance
of the tests, and estimates of sensitivity, specificity, and
predictive values were calculated and reported with 95 %
confidence intervals (CI). All other results in this report are
deemed statistically significant at a nominal 5 % level
(P \ 0.05) without adjustment for multiple comparisons.

Results

A total of 54 patients diagnosed with HER2-positive early- Fig. 1 Diagram depicting identification and enrollment of study
stage breast cancer and eligible for trastuzumab adjuvant participants

123
294 Breast Cancer Res Treat (2012) 134:291–298

Table 1 Distribution of age and BMI and number of tests performed by LVEF group
Normal LVEF (N = 35) Decreased LVEF (N = 14)
Mean Median Mean Median P value

Age 55.41 54.9 59.58 57.05 0.4128

BMI 30.17 27.73 29.58 29.80 0.7567
Number of ECHO/MUGA tests 4.5 5.0 3.4 3.0 0.0300
Number of laboratory tests of CRP 11.3 13.0 8.4 9.0 0.0428
Number of laboratory tests of BNP 12.0 13.0 8.5 9.0 0.0256
BMI body mass index, LVEF left ventricular ejection fraction, ECHO/MUGA echocardiogram/multiple gated acquisition, CRP C-reactive
protein, BNP B-type natriuretic peptide

Table 2 Distribution of known cardiac risk factors by LVEF group in 9 subjects (64.3 %). Trastuzumab was discontinued or
held in 5/14 patients that experienced a clinically significant
Risk factor Normal Decreased P value
LVEF LVEF
decrease in LVEF. Following treatment plan modification,
N (%) N (%) hs-CRP levels improved in all 5 subjects.

Neoadjuvant chemotherapy 9 (27.5 %) 3 (21.4 %) 1.0000

Radiation therapy 15 (42.9 %) 5 (35.7%) 0.6458
Anthracycline 15 (42.9 %) 9 (64.3 %) 0.1752 Discussion
Taxane 32 (91.4 %) 14 (100 %) 0.5479
Endocrine therapy 19 (54.3 %) 4 (28.6 %) 0.1722
Cardiotoxicity, as defined by decrease in LVEF, has been
Hypertension 13 (37.1 %) 4 (28.6 %) 0.7430
observed following trastuzumab adjuvant treatment com-
bined with standard chemotherapy in as many as 27–34 %
Diabetes 4 (11.4 %) 2 (14.3 %) 1.0000
of patients [5, 9, 20]. Consistent with these findings, 28 %
Hypercholesterolemia 13 (37.1 %) 4 (28.6 %) 0.7430
(14/49) of women enrolled in this study who were treated
Coronary artery disease 2 (5.71 %) 3 (21.4 %) 0.1330
with trastuzumab experienced a clinically significant
Smoking 20 (57.1 %) 11 (78.6 %) 0.2024
decrease in LVEF. Rates of cardiotoxicity are somewhat
HRT/OCP 7 (20.0 %) 1 (7.14 %) 0.4133
lower in newer adjuvant trials of trastuzumab, but likely
Acetylsalicylic acid 3 (8.57 %) 3 (21.4 %) 0.3386
underestimate risk due to careful selection of patients to
Statin 8 (25.7 %) 3 (21.4 %) 1.0000
minimize cardiac events by excluding patients with
Antihypertensives 11 (31.4 %) 5 (35.7 %) 0.8223
underlying cardiac disease or abnormal baseline or post-
Cancer stage 0.1333
anthracycline treatment LVEF [5]. Such patients were not
I 18 (51.4 %) 3 (21.4 %)
excluded from this study. The importance of detecting and
II 11 (31.4 %) 7 (50.0 %)
preventing the development of significant cardiotoxicity is
III 6 (17.1 %) 4 (28.6 %) emphasized by the finding that in a cohort of 5-year breast
LVEF left ventricular ejection fraction, HRT/OCP hormone replace- cancer survivors, cardiovascular disease at baseline was
ment therapy/oral contraceptive pills associated with a significant increase in the risk of death
from other, non-breast cancer causes, accounting for 60 %
Test characteristics of hs-CRP at various cut-offs are of deaths in this population [21].
shown in Table 4. At the best cutoff ([3 mg/L), testing for Studies have demonstrated the utility of BNP, hs-CRP,
abnormal hs-CRP has a sensitivity of 92.9 % (95 % CI and cTnI in predicting cardiovascular disease in a number of
66.1–99.8) and specificity of 45.7 % (95 % CI 28.8–63.4) settings. In this study, only elevated hs-CRP, not BNP or
for detecting a clinically significant decrease in LVEF in cTnI, was associated with increased risk for a clinically
breast cancer patients treated with trastuzumab. The nega- significant decrease in LVEF. In 2000, Ridker et al. [22]
tive predictive value is 94.1 % (95 % CI 70.3–99.9), and the found elevated hs-CRP to be predictive of cardiovascular
positive predictive value is 40.6 % (95 % CI 23.7–59.4). In disease in women. Elevated hs-CRP has since been associ-
the 14 subjects with cardiotoxicity, maximum hs-CRP was ated with poor prognostic indicators, such as decreased
measured a median of 77.5 days before toxicity was iden- LVEF and diastolic dysfunction, in patients with a variety of
tified via decrease in LVEF (range 14–359 days). Most cardiovascular maladies including acute myocardial infarc-
subjects (8/14) had a peak level within 90 days of toxicity. tion [23], chronic stable angina [24], left ventricular systolic
In these 14 subjects with cardiotoxicity, the last hs-CRP heart failure [17], idiopathic left ventricular dysfunction
measurement before toxicity was still abnormal ([3 mg/L) [16], and congestive heart failure [15, 25]. The results

123
Breast Cancer Res Treat (2012) 134:291–298 295

Fig. 2 Dot diagram depicting

the distribution of maximum
pre-toxicity a hs-CRP values in
each outcome group and b BNP
values in each outcome group

was more accurate in predicting cardiovascular events in

men with type 2 diabetes mellitus than either marker alone.
The combination of hs-CRP and BNP as a biomarker of
cardiovascular disease has also been found to be useful in the
management of patients with stable coronary heart disease
[27] and dilated cardiomyopathy [28]. In contrast, in this
study, when the presence of abnormal hs-CRP or BNP was
assessed as a marker of cardiotoxicity, there was no differ-
ence between patients with normal or decreased LVEF,
suggesting that BNP would not make a good additional
biomarker in this particular patient population.
A number of preliminary studies have demonstrated the
value of BNP and cTnI in predicting cardiotoxicity. How-
ever, studies showing an association between elevated BNP
and trastuzumab-induced cardiotoxicity considered only
baseline values [29, 30]. Those that examined BNP levels
over time found no association [31, 32], in agreement with
the present findings. Results for cTnI are mixed. Kutteh et al.
[29] observed association between baseline cTnI levels and
Fig. 3 Receiver operating characteristic (ROC) curve analysis of
cardiotoxicity. Sawaya et al. [32] observed association
maximum hs-CRP and maximum BNP pre-toxicity in patients with
normal and decreased LVEF. When the test cannot distinguish between cTnI levels at 3 months post-treatment initiation
between positive or negative on each standard, i.e., where there is no and development of cardiotoxicity at 6 months. Similarly,
difference between the two distributions, the area will be equal to 0.5. Cardinale et al. [33] found elevated cTnI levels to correlate
When there is a perfect separation of the values of the two groups, i.e.
with risk for irreversible cardiotoxicity in patients who
there no overlapping of the distributions, the area under the ROC
curve equals 1 received trastuzumab after treatment with an anthracycline,
but not in those who had not been treated with an anthracy-
presented here suggest that elevated serum hs-CRP may also cline. In contrast, neither Goel et al. [30] nor Fallah-Rad et al.
be useful as a biomarker for early detection of risk for car- [31] detected changes in cTnI with trastuzumab treatment,
diotoxicity in breast cancer patients undergoing adjuvant consistent with our findings. Fallah-Rad et al. [31] also
trastuzumab treatment combined with standard chemother- assessed CRP and found no association with cardiotoxicity;
apy. Interestingly, Tsuruda et al. [26] found that using a however, they did not use the high-sensitivity test employed
combination of abnormal BNP and hs-CRP as biomarkers in this study. A recent study by Morris et al. [34] examined

123
296 Breast Cancer Res Treat (2012) 134:291–298

Table 3 Association between decreased LVEF, hs-CRP, CAD, and anthracycline use via multivariable logistic regression
B SE P value OR 95 % CI for OR
Lower Upper

Anthracycline use (Y) 1.302 .897 .147 3.677 .634 21.321

CAD present (Y) 3.054 1.441 .034 21.194 1.257 357.332
Any abnormal hs-CRP (Y) 2.634 1.258 .036 13.926 1.184 163.766
OR odds ratio, LVEF left ventricular ejection fraction, CAD coronary artery disease, hs-CRP high-sensitivity C-reactive protein, B beta coef-
ficient, SE standard error

Table 4 Criterion values and coordinates of the ROC curve for hs-CRP
Hs-CRP criterion (mg/L) Sensitivity 95 % CI Specificity 95 % CI ?PV 95 % CI -PV 95 % CI

[1 92.86 66.1–99.8 14.29 4.8–30.3 30.2 17.2–46.1 83.3 31.1–99.8

[3 92.86 66.1–99.8 45.71 28.8–63.4 40.6 23.7–59.4 94.1 70.3–99.9
[6 78.57 49.2–95.3 60.00 42.1–76.1 44.0 24.4–65.1 87.5 67.6–97.3
[9 57.14 28.9–82.3 74.29 56.7–87.5 47.1 22.3–72.9 81.2 63.6–92.8
ROC receiver operating characteristic, hs-CRP high-sensitivity C-reactive protein, PV predictive value

serum biomarkers, including hs-CRP, for the early detection early events such as surgery, anthracycline use or other
of cardiotoxicity in patients with breast cancer treated with chemotherapy, left-side radiation therapy, and stressful
trastuzumab in combination with an anthracycline-based events early in the course of breast cancer diagnosis may
chemotherapy regimen. Abnormal hs-CRP levels were result in the elevation of acute phase reactants, such as
observed but did not correlate with asymptomatic declines in CRP. However, our data suggest that subjects with normal
LVEF. A major limitation to the study by Morris et al. [34] hs-CRP levels (\ 3 mg/L) are unlikely to develop trast-
was timing of the blood draw, which occurred immediately uzumab-induced cardiotoxicity (94.1 % negative predic-
before treatment, and was not timed with respect to trast- tive value). Thus, hs-CRP may be a useful marker in
uzumab treatment, but rather to the administration of che- decision making regarding who may not need stringent
motherapy. Furthermore, patients with underlying cardiac cardiac follow-up. Those with elevated hs-CRP levels
dysfunction, potentially those most at risk for cardiotoxicity, should continue to receive stringent monitoring as per
were excluded, and all patients enrolled in the study were current standards of care.
required to adhere to a single chemotherapeutic protocol, Potential markers of cardiotoxicity such as BNP and
resulting in a high drop-out rate [34]. In this study, the con- cTnI tend to be specific for cardiomyopathy due to local-
sistent measurement of hs-CRP levels at initiation and ization of these proteins within the heart. In contrast,
administration of each subsequent dose of trastuzumab as hepatocytes are the source of CRP, which is produced
well as regular assessment of LVEF by ECHO or MUGA during many inflammatory conditions such as infection,
every 3–4 months during treatment per standard institutional and is therefore not specific to cardiomyopathy. Thus, the
practice provides a more complete picture of how hs-CRP relatively low specificity and positive predictive value of
levels relate to trastuzumab-induced cardiotoxicity as hs-CRP for decreased LVEF are not particularly surprising.
defined by a clinically significant decrease in LVEF. However, a false positive (elevated hs-CRP in the absence
The use of elevated hs-CRP as a biomarker to predict of cardiotoxicity) would have a relatively low associated
trastuzumab-induced cardiotoxicity appears highly sensi- cost in breast cancer patients undergoing adjuvant trast-
tive (92.9 %) with a high negative predictive value uzumab therapy with elevated hs-CRP indicating need for
(94.1 %), suggesting that normal hs-CRP levels may be continuation of routine observation during trastuzumab
associated with low future risk for decreased LVEF. treatment to ensure cardiac health and prevent irreversible
However, because elevated hs-CRP has a relatively low damage. In this respect, abnormal hs-CRP levels were
specificity (45.7 %) and positive predictive value (40.6 %) found a median of 78 days before detection of cardiotox-
for decreased LVEF with a high rate of false positives, it icity by decreased LVEF and thus, normal levels imply that
does not reliably imply toxicity. This is expected because there is unlikely to be immediate concerns regarding

123
Breast Cancer Res Treat (2012) 134:291–298 297

cardiotoxicity as measured by standard ECHO or MUGA References

testing.
A major limitation of this study is the use of decreased 1. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Baja-
LVEF as a surrogate for cardiotoxicity. In the United monde A, Fleming T, Eiermann W, Wolter J, Pegram M, Baselga
J, Norton L (2001) Use of chemotherapy plus a monoclonal
States, *50 % of patients diagnosed with heart failure
antibody against HER2 for metastatic breast cancer that overex-
actually have decreased LVEF [35–37], and the overall presses HER2. N Engl J Med 344:783–792
cardiac outcomes of heart failure patients with and without 2. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire
decreased LVEF are similar [37]. At present, the long-term WL (1987) Human breast cancer: correlation of relapse and
survival with amplification of the HER-2/neu oncogene. Science
implications of asymptomatic changes in LVEF are unclear
235:177–182
[5, 6]; therefore, it is possible that increased hs-CRP may 3. Costa RB, Kurra G, Greenberg L, Geyer CE (2010) Efficacy and
be a marker of trastuzumab-induced cardiotoxicity even cardiac safety of adjuvant trastuzumab-based chemotherapy
when decreased LVEF is not detected. This study suggests regimens for HER2-positive early breast cancer. Ann Oncol
21:2153–2160
that abnormal hs-CRP has a high sensitivity (92.9 %) and
4. Viani GA, Afonso SL, Stefano EJ, De Fendi LI, Soares FV
negative predictive value (94.4 %) for decreased LVEF, (2007) Adjuvant trastuzumab in the treatment of her-2-positive
but the specificity (48.6 %) and positive predictive value early breast cancer: a meta-analysis of published randomized
(41.9 %) are low. However, if assessment of decreased trials. BMC Cancer 7:153
5. Chien AJ, Rugo HS (2010) The cardiac safety of trastuzumab in
LVEF is only capable of identifying approximately half of
the treatment of breast cancer. Expert Opin Drug Saf 9:335–346
the patients experiencing cardiotoxicity, as suggested by 6. Routledge HC, Rea DW, Steeds RP (2006) Monitoring the intro-
Gottdeiner et al. [37], then abnormal hs-CRP may truly duction of new drugs—Herceptin to cardiotoxicity. Clin Med
have a substantially greater specificity and positive pre- 6:478–481
7. Tripathy D, Slamon DJ, Cobleigh M, Arnold A, Saleh M, Mor-
dictive value for trastuzumab-induced cardiotoxicity.
timer JE, Murphy M, Stewart SJ (2004) Safety of treatment of
This study is also limited by the relatively small sample metastatic breast cancer with trastuzumab beyond disease pro-
size inherent to such an exploratory study. In addition, data gression. J Clin Oncol 22:1063–1070
analysis does not include certain risk factors for cardio- 8. Ewer MS, Vooletich MT, Durand JB, Woods ML, Davis JR,
Valero V, Lenihan DJ (2005) Reversibility of trastuzumab-rela-
vascular disease, including use of non-steroidal anti-
ted cardiotoxicity: new insights based on clinical course and
inflammatory drugs and corticosteroids, as very few response to medical treatment. J Clin Oncol 23:7820–7826
patients were being treated with these medications. Finally, 9. Seidman A, Hudis C, Pierri MK, Shak S, Paton V, Ashby M,
because patients were not required to adhere to a single Murphy M, Stewart SJ, Keefe D (2002) Cardiac dysfunction in the
trastuzumab clinical trials experience. J Clin Oncol 20:1215–1221
treatment protocol, there is considerable heterogeneity in
10. Kazanegra R, Cheng V, Garcia A, Krishnaswamy P, Gardetto N,
the patient population studied, which may mask some of Clopton P, Maisel A (2001) A rapid test for B-type natriuretuc
the effects of trastuzumab treatment, but may also increase peptide correlates with falling wedge pressures in patients treated
generalizability of the results. Study generalizability may for decompensated heart failure: a pilot study. J Card Fail 7:21–29
11. Krishnaswamy P, Lubien E, Clopton P, Koon J, Kazanegra R,
also benefit from the minimal interference with standard
Wanner R, Gardetto N, Garcia A, DeMaria A, Maisel AS (2001)
clinical practice and flexibility allowed to the managing Utility of B-natriuretic peptide levels in identifying patients with
oncologists, suggesting that hs-CRP testing may be rela- left ventricular systolic or diastolic dysfunction. Am J Med
tively easy to incorporate into clinical practice. 111:274–279
12. Cardinale D, Sandri MT, Martinoni A, Borghini E, Civelli M, La-
As a biomarker for identifying the women with early-stage
mantia G, Cinieri S, Martinelli G, Fiorentini C, Cipolla CM (2002)
breast cancer who are at low risk for asymptomatic trast- Myocardial injury revealed by plasma troponin I in breast cancer
uzumab-induced cardiotoxicity, hs-CRP testing holds prom- treated with high-dose chemotherapy. Ann Oncol 13:710–715
ise for clinical use. To our knowledge, this is the first study 13. Lubien E, DeMaria A, Krishnaswamy P, Copton P, Koon J,
Kazanegra R, Gardetto N, Wanner E, Maisel AS (2002) Utility of
documenting the utility of a biomarker that is less expensive,
B-natriuretic peptide in detecting diastolic dysfunction: compar-
reproducible, and easily obtainable with rapid results for ison with Doppler velocity recordings. Circulation 105:595–601
evaluating cardiotoxicity related to trastuzumab therapy. We 14. Sandri MT, Cardinale D, Zorzino L, Passerini R, Lentati P,
recommend further evaluation in a larger setting. Martinoni A, Martinenlli G, Cipolla CM (2003) Minor increases
in plasma troponin I predict decreased left ventricular ejection
fraction after high-dose chemotherapy. Clin Chem 49:248–252
Acknowledgments The authors thank Marie Fleisner of the
15. Xue C, Feng Y, Wo J, Li Y (2006) Prognostic value of high-
Marshfield Clinic Research Foundation’s Office of Scientific Writing
sensitivity C-reactive protein in patients with chronic heart fail-
and Publication for proofreading/copyediting assistance. Funding for
ure. N Z Med J 119:U2314
this study was provided by the Marshfield Clinic Physician Research
16. Giannessi D, Colotti C, Maltinti M, Del Ry S, Prontera C, Turchi
Funds and the Marshfield Clinic Research Foundation Disease Spe-
S, Labbate A, Neglia D (2007) Circulating heat shock proteins
cific Funds.
and inflammatory markers in patients with idiopathic left ven-
Conflicts of interest The authors declare no potential conflicts of tricular dysfunction: their relationships with myocardial and
interest. microvascular impairment. Cell Stress Chaperones 12:265–274

123
298 Breast Cancer Res Treat (2012) 134:291–298

17. Tang WH, Shrestha K, Van Lente F, Troughton RW, Martin MG, 28. Ishikawa C, Tsutamoto T, Fujii M, Sakai H, Tanaka T, Horie M
Borowski AG, Jasper S, Klein AL (2008) Usefulness of C-reac- (2006) Prediction of mortality by high-sensitivity C-reactive
tive protein and left ventricular diastolic performance for prog- protein and brain natriuretic peptide in patients with dilated
nosis in patients with left ventricular systolic heart failure. Am J cardiomyopathy. Circ J 70:857–863
Cardiol 101:370–373 29. Kutteh LA, Hobday T, Jaffe A, LaPlant B, Hillman D, Kaufman
18. Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon P, Davidson N, Martino S, Moreno A, Perez E, Central Cancer
RO III, Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Treatment Group (2007) A correlative study of cardiac bio-
Rifai N, Smith SC Jr, Taubert K, Tracy RP, Vinicor F, Centers for markers and left ventricular ejection fraction (LVEF) from
Disease Control and Prevention; American Heart Association N9831, a phase III randomized trial of chemotherapy and trast-
(2003) Markers of inflammation and cardiovascular disease: uzumab as adjuvant therapy for HER2-positive breast cancer.
application to clinical and public health practice: a statement for J Clin Oncol 25(18):579
healthcare professionals from the Centers for Disease Control and 30. Goel S, Simes RJ, Beith JM (2011) Exploratory analysis of car-
Prevention and the American Heart Association. Circulation diac biomarkers in women with normal cardiac function receiv-
107:499–511 ing trastuzumab for breast cancer. Asia Pac J Clin Oncol 7:
19. Heeschen C, Hamm CW, Goldmann B, Deu A, Langenbrink L, 276–280
White HD (1999) Troponin concentrations for stratification of 31. Fallah-Rad N, Walker JR, Wassef A, Lytwyn M, Bohonis S, Fang
patients with acute coronary syndromes in relation to therapeutic T, Tian G, Kirkpatrick ID, Singal PK, Krahn M, Grenier D, Jassal
efficacy of tirofiban. PRISM Study Investigators. Platelet DS (2011) The utility of cardiac biomarkers, tissue velocity and
Receptor Inhibition in Ischemic Syndrome Management. Lancet strain imaging, and cardiac magnetic resonance imaging in pre-
354:1757–1762 dicting early left ventricular dysfunction in patients with human
20. Tan-Chiu E, Yothers G, Romond E, Geyer CE Jr, Ewer M, Keefe epidermal growth factor receptor II-positive breast cancer treated
D, Shannon RP, Swain SM, Brown A, Fehrenbacher L, Vogel VG, with trastuzumab therapy. J Am Coll Cardiol 57:2263–2270
Seay TE, Rastogi P, Mamounas EP, Wolmark N, Bryant J (2005) 32. Sawaya H, Sebag IA, Plana JC, Januzzi JL, Ky B, Cohen V,
Assessment of cardiac dysfunction in a randomized trial com- Gosavi S, Carver JR, Wiegers SE, Martin RP, Picard MH,
paring doxorubicin and cyclophosphamide followed by paclitaxel, Gerszten RE, Halpern EF, Passeri J, Kuter I, Scherrer-Crosbie M
with or without trastuzumab as adjuvant therapy in node-positive, (2011) Early detection and prediction of cardiotoxicity in che-
human epidermal growth factor receptor 2-overexpressing breast motherapy-treated patients. Am J Cardiol 107:1375–1380
cancer: NSABP B-31. J Clin Oncol 23:7811–7819 33. Cardinale D, Colombo A, Torrisi R, Sandri MT, Civelli M,
21. Chapman JA, Meng D, Shepherd L, Parulekar W, Ingle JN, Muss Salvatici M, Lamantia G, Colombo N, Cortinovis S, Dessanai
HB, Palmer M, Yu C, Goss PE (2008) Competing causes of death MA, Nolè F, Veglia F, Cipolla CM (2010) Trastuzumab-induced
from a randomized trial of extended adjuvant endocrine therapy cardiotoxicity: clinical and prognostic implications of troponin I
for breast cancer. J Natl Cancer Inst 100:252–260 evaluation. J Clin Oncol 28:3910–3916
22. Ridker PM, Hennekens CH, Buring JE, Rifai N (2000) C-reactive 34. Morris PG, Chen C, Steingart R, Fleisher M, Lin N, Moy B, Come
protein and other markers of inflammation in the prediction of S, Sugarman S, Abbruzzi A, Lehman R, Patil S, Dickler M,
cardiovascular disease in women. N Engl J Med 342:836–843 McArthur HL, Winer E, Norton L, Hudis CA, Dang CT (2011)
23. Arruda-Olson AM, Enriquez-Sarano M, Bursi F, Weston SA, Troponin I and C-reactive protein are commonly detected in
Jaffe AS, Killian JM, Roger VL (2010) Left ventricular function patients with breast cancer treated with dose-dense chemotherapy
and C-reactive protein levels in acute myocardial infarction. Am J incorporating trastuzumab and lapatinib. Clin Cancer Res 17:
Cardiol 105:917–921 3490–3499
24. Arroyo-Espiliguero R, Avanzas P, Quiles J, Kaski JC (2009) 35. Mosterd A, Hoes AW, de Bruyne MC, Deckers JW, Linker DT,
Predictive value of coronary artery stenoses and C-reactive pro- Hofman A, Grobbee DE (1999) Prevalence of heart failure and
tein levels in patients with stable coronary artery disease. Ath- left ventricular dysfunction in the general population: The Rot-
erosclerosis 204:239–243 terdam Study. Eur Heart J 20:447–455
25. Windram JD, Loh PH, Rigby AS, Hanning I, Clark AL, Cleland 36. Vasan RS, Larson MG, Benjamin EJ, Evans JC, Reiss CK, Levy
JG (2007) Relationship of high-sensitivity C-reactive protein to D (1999) Congestive heart failure in subjects with normal versus
prognosis and other prognostic markers in outpatients with heart reduced left ventricular ejection fraction: prevalence and mor-
failure. Am Heart J 153:1048–1055 tality in a population-based cohort. J Am Coll Cardiol 33:1948–
26. Tsuruda T, Kato J, Sumi T, Mishima K, Masuyama H, Nakao H, 1955
Imamura T, Eto T, Kitamura K (2007) Combined use of brain 37. Gottdiener JS, McClelland RL, Marshall R, Shemanski L, Fur-
natriuretic peptide and C-reactive protein for predicting cardio- berg CD, Kitzman DW, Cushman M, Polak J, Gardin JM, Gersh
vascular risk in outpatients with type 2 diabetes mellitus. Vasc BJ, Aurigemma GP, Manolio TA (2002) Outcome of congestive
Health Risk Manag 3:417–423 heart failure in elderly persons: influence of left ventricular sys-
27. Ndrepepa G, Kastrati A, Braun S, Mehilli J, Niemöller K, von tolic function. The Cardiovascular Health Study. Ann Intern Med
Beckerath N, von Beckerath O, Vogt W, Schömig A (2006) 137:631–639
N-terminal probrain natriuretic peptide and C-reactive protein in
stable coronary heart disease. Am J Med 119(355):e1–e8

123