You are on page 1of 7

CLB-09486; No.

of pages: 7; 4C:
Clinical Biochemistry xxx (2017) xxxxxx

Contents lists available at ScienceDirect

Clinical Biochemistry

journal homepage: www.elsevier.com/locate/clinbiochem

NT-proBNP is increased in differentiated thyroid carcinoma patients and


may predict cardiovascular risk
Esther N. Klein Hesselink a,b, Anouk N.A. van der Horst-Schrivers b, Iwan C.C. van der Horst c,
Stephan J.L. Bakker d, Anneke C. Muller Kobold e, Adrienne H. Brouwers f, Geertruida H. de Bock g,
Jourik A. Gietema h, Robin P.F. Dullaart b, Thera P. Links b,, Joop D. Lefrandt a
a
University of Groningen, University Medical Center Groningen, Department of Vascular Medicine, Groningen, The Netherlands
b
University of Groningen, University Medical Center Groningen, Department of Endocrinology, Groningen, The Netherlands
c
University of Groningen, University Medical Center Groningen, Department of Critical Care, Groningen, The Netherlands
d
University of Groningen, University Medical Center Groningen, Department of Nephrology, Groningen, The Netherlands
e
University of Groningen, University Medical Center Groningen, Department of Laboratory Medicine, Groningen, The Netherlands
f
University of Groningen, University Medical Center Groningen, Department of Nuclear Medicine and Molecular Imaging, Groningen, The Netherlands
g
University of Groningen, University Medical Center Groningen, Department of Epidemiology, Groningen, The Netherlands
h
University of Groningen, University Medical Center Groningen, Department of Medical Oncology, Groningen, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: Chronic suppression of TSH in patients treated for differentiated thyroid carcinoma (DTC) may
Received 29 November 2016 induce cardiac damage and increase risk for cardiovascular events and premature mortality. We aimed to com-
Received in revised form 27 January 2017 pare circulating concentrations of N-terminal pro Brain Natriuretic Peptide (NT-proBNP) of DTC patients with
Accepted 23 February 2017 controls, and to investigate whether higher NT-proBNP is associated with an increased risk for cardiovascular
Available online xxxx
events and all-cause mortality in DTC patients.
Methods: Serum NT-proBNP levels were determined in 266 DTC patients, median 10.4 [IQR 4.118.5] years
Keywords:
NT-proBNP
after DTC diagnosis, and compared to 798 age- and sex-matched controls. Using multivariable Cox regression
Differentiated thyroid cancer analyses, the association of NT-proBNP with cardiovascular events and all-cause mortality was determined. Haz-
Cardiovascular risk ard ratios (HR) and 95% condence intervals (CIs) were expressed per SD increase of log-transformed NT-
proBNP.
Results: Mean age SD of DTC patients and controls was 54.8 14.5 and 54.8 12.8 years, respectively;
74% were women. Median NT-proBNP level was 70 [40119] ng/L for DTC patients vs. 49 [2589] ng/L for con-
trols (p b 0.001). During median follow-up of 8.6 [6.69.0] years, 30 DTC patients (11.4%) had a cardiovascular
event and 38 (14.4%) died. Higher NT-proBNP was associated with an increased risk for cardiovascular events
and all-cause mortality, age- and sex-adjusted HRs (95% CIs) 3.22 (2.174.79) and 1.61 (1.172.23), respectively.
In further models with adjustment for cardiovascular risk factors, NT-proBNP remained independently associat-
ed with outcome.
Conclusion: NT-proBNP levels are elevated in patients with DTC, and are associated with an increased risk for
cardiovascular events and all-cause mortality. Determination of NT-proBNP may identify DTC patients at in-
creased cardiovascular risk, who could benet from more stringent cardiovascular risk surveillance.
2017 Published by Elsevier Inc. on behalf of The Canadian Society of Clinical Chemists.

1. Introduction related adverse cardiovascular effects, such as atrial brillation, diastolic


and systolic cardiac dysfunction, increased left ventricular mass and a
Differentiated thyroid carcinoma (DTC) is a malignancy with a high prothrombotic prole [25]. Moreover, DTC patients have an increased
10-year relative survival of 95% [1]. Therefore, there is an increasing in- risk for cardiovascular mortality, which may in part be attributable to
terest for long-term adverse effects of treatment in patients with DTC. thyroid hormone suppression therapy (THST) [6].
Recent evidence shows that DTC survivors are at risk for treatment- Brain natriuretic peptide (BNP) has been shown to be a strong pre-
dictor of cardiovascular events and mortality [7]. BNP is produced in
the heart in response to cardiac wall stress, induced by volume excess
Corresponding author at: University of Groningen, University Medical Center
Groningen, Department of Endocrinology, HPC AA31, P.O. Box 30.001, 9700 RB
or pressure overload [7]. In addition, chemical, ischemic and inamma-
Groningen, The Netherlands. tory factors can affect BNP production [8]. BNP protects the cardiovascu-
E-mail address: t.p.links@umcg.nl (T.P. Links). lar system from the effects of volume overload and counteracts

http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
0009-9120/ 2017 Published by Elsevier Inc. on behalf of The Canadian Society of Clinical Chemists.

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
2 E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx

hemodynamic stress, endothelial dysfunction and vascular remodeling, up at the UMCG. These samples were considered residual material.
as it lowers blood pressure, causes vasodilation, natriuresis, and diure- The serum sample from the earliest date within the period 20042008
sis, and inhibits myocardial and vascular inammation [9]. To form per DTC patient was selected, and NT-proBNP was measured in
BNP, pro-BNP is split, with N-terminal proBNP (NT-proBNP) being re- these residual samples. For controls, NT-proBNP was measured in
leased as an inactive fragment [7]. In vitro, NT-proBNP is more stable baseline samples, stored at 80 C, obtained from PREVEND partici-
than BNP, and represents a predictor for cardiovascular events and all- pants at the rst screening round in 1997/1998 [16]. Samples were
cause mortality in the general population as well as in several patient thawed at room temperature, and subsequently the NT-proBNP mea-
categories [1013]. surements were performed on the Elecsys 2010, making use of an
Relatively high BNP levels have been reported in hyperthyroidism electrochemiluminescent sandwich immunoassay (Elecsys proBNP,
[14,15], which has been proposed to be consequent to a stimulatory ef- Roche Diagnostics, Mannheim, Germany). The analytical range was
fect of thyroid hormone on BNP gene transcription itself [15], and possi- 535,000 ng/L, intra- and inter-assay inaccuracy were 1.21.5% and
bly also to adverse cardiovascular effects of an excess of thyroid 4.45.0%, respectively [21].
hormone [14]. These ndings support the rationale to specically assess
this biomarker in DTC patients, who are intentionally in a state of sub-
clinical hyperthyroidism due to THST in at least a part of follow-up. Fur- 1.1.3. Study denitions
thermore, the value of biomarkers such as NT-proBNP to assess Baseline was dened as the date of serum sampling for NT-proBNP
cardiovascular prognosis have not been elucidated in DTC patients, measurement. Survival time was dened as the time from baseline to
whereas it may be of importance to identify patients at increased risk the date a DTC patient had a cardiovascular event, died, was lost to fol-
in time to start cardiovascular protective measures. low-up, or to the census date, which was May 1, 2015. The baseline car-
Our aims were 1) to determine whether NT-proBNP is elevated in diovascular risk factors consisted of age, sex, smoking, body mass index
DTC patients, and 2) to evaluate whether NT-proBNP could serve as a (BMI), presence of diabetes mellitus (DM), hypertension, and hyper-
prognostic marker of cardiovascular risk and all-cause mortality in this cholesterolemia, estimated glomerular ltration rate (eGFR), and a
patient category during follow-up. known history of cardiovascular disease. DM, hypertension and hyper-
cholesterolemia were dened as the use of prescription drugs to treat
1.1. Patient and methods these conditions. eGFR was calculated with the Modication of Diet in
Renal Disease (MDRD) equation using the baseline creatinine level, or,
1.1.1. Study design and population when not available, the closest available level within 2 years, or a
In this cohort study with prospectively identied patients, all pa- more recent level when the patient was in a clinically stable state. A his-
tients in follow-up for DTC (including both papillary and follicular thy- tory of cardiovascular disease was dened as documented acute coro-
roid carcinoma) at the University Medical Center Groningen (UMCG), nary syndrome (ACS), coronary artery disease, percutaneous coronary
The Netherlands, were eligible. Criteria for inclusion were: DTC diagno- intervention (PCI), coronary artery bypass graft (CABG), abdominal aor-
sis between January 1, 1961 and January 1, 2006, age at diagnosis of tic aneurysm (AAA), intervention for peripheral artery disease (PAD),
18 years or higher, treatment with total thyroidectomy and radioiodine heart failure, a cerebrovascular accident (CVA), or a transient ischemic
ablation, DTC survival of at least 1 year, and availability of at least 1 fol- attack (TIA) prior to the NT-proBNP measurement. In case of missing
low-up record and a serum sample stored at 80 C collected between characteristics, DTC patients were contacted for further information re-
2004 and 2008. If NT-proBNP measurement failed, patients were ex- garding the particular characteristic.
cluded. Furthermore, we excluded patients who had a cardiovascular At the end of follow-up, medical records of DTC patients were
event, or died within 6 weeks after serum sampling for NT-proBNP mea- reassessed to determine cardiovascular events and causes of death. Pri-
surement for the analysis of NT-proBNP as a prognostic marker, in order mary endpoint was the combined endpoint of cardiovascular events,
to ensure that NT-proBNP was not measured during the initial phase of which was dened as occurrence of the following fatal or non-fatal
a (cardiovascular) event. events during follow-up: a cardiac event (ACS, PCI, CABG), cerebrovas-
For each DTC patient we selected three age- and sex-matched con- cular event (CVA, TIA), AAA (ruptured or requiring intervention) or in-
trols, to be able to compare the NT-proBNP levels between DTC patients tervention for PAD or new-onset heart failure. Heart failure was scored
and a sample representative of the general population. Controls were by a blinded panel consisting of a cardiologist (I.C.C.H.) and an internist
selected from the general population sample of the prospective popula- specialized in vascular medicine (J.D.L). The nal diagnosis heart failure
tion based Prevention of REnal and Vascular ENd-stage Disease was established by review of medical history, physical examinations,
(PREVEND) study, as previously described [16]. echocardiography ndings and nuclear imaging results without knowl-
Treatment and follow-up of patients with DTC was according to our edge of the cardiac biomarkers. Secondary endpoint was all-cause mor-
institutional protocol [17,18], and the Dutch guideline [19]. Treatment tality. Death causes of category I1099 of the International Classication
consisted of a (near) total thyroidectomy and a central and/or lateral of Disease version 10 were dened as cardiovascular deaths. Medical re-
neck dissection if necessary. Postoperatively, patients were classied ac- cords and, if available, autopsy reports were assessed to determine
cording to the risk of DTC recurrence using the TNM classication [20], cause of death. Furthermore, the general practitioner was contacted
as low (Tx-T2, Nx-N0, Mx-M0), intermediate (any T3 or N1) or high when the cause of death was not available from the medical record.
(any T4 or M1 tumor) risk (later referred to as DTC risk classication),
and underwent radioiodine ablation therapy. Thereafter, patients
started taking thyroid hormones for replacement or TSH suppression 1.2. Statistics
therapy, since TSH is regarded a growth factor for DTC cells. The TSH tar-
get level varied over time and between low, intermediate and high risk Data were presented as mean standard deviation (SD), or median
patients, as previously described [6]. Anonymity was assured using and interquartile range (IQR). NT-proBNP values were compared be-
unique codes. The institutional review board of the UMCG declared tween DTC patients and controls using the Mann-Whitney U test. Differ-
that based on Dutch law no further approval for use of residual serum ences in baseline characteristics of DTC patients between quartiles of
was needed (METc 2012/161). NT-proBNP were tested by ANOVA, Kruskal Wallis test, Chi-square or
Fisher exact test, as appropriate. Missing BMI data (of 16 DTC patients)
1.1.2. NT-proBNP measurements were imputed based on a regression including age and sex; missing
Since 2004, the serum of redundant blood that was drawn during eGFR data (of 7 patients) were imputed based on age, sex, hypertension,
regular blood testing was stored at 80 C for DTC patients in follow- DM, and history of cardiovascular disease (n = 5 imputations) [22].

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx 3

The Kaplan-Meier method was used to estimate time-to-event All tests were two-sided; a p-value of b0.05 was considered statisti-
curves for the endpoints cardiovascular events and all-cause mortality cally signicant. Software package IBM SPSS version 22.0 (IBM, Armonk,
by NT-proBNP quartiles, with a subsequent log-rank test. NY), and STATA version 14.0 (StataCorp, College Station, TX) were used.
Cox proportional hazards analyses were used to test NT-proBNP as a
prognostic marker for cardiovascular events and all-cause mortality. 2. Results
NT-proBNP was logarithmically transformed and evaluated in the Cox
regression models per 1 SD increase. Both crude (model 1) and several 2.1. NT-proBNP in patients and controls
adjusted analyses (models 2 to 9) were performed. Model 2 was adjust-
ed for age and sex. In order to avoid inclusion of too many variables NT-proBNP levels of 266 DTC patients and 798 age- and sex-
given the number of outcomes, further models (models 3 to 8) were matched controls were included in the analysis (Fig. 1). At baseline,
made with additional adjustments to model 2. These models were ad- mean SD age was 54.8 14.5 years for DTC patients, and 54.8
justed for hypercholesterolemia, hypertension, DM, BMI, eGFR, and 12.8 years for controls. A total of 196 patients (74%), and 588 controls
tumor histology and DTC risk classication, respectively, in addition to (74%) were female. Baseline characteristics of DTC patients by NT-
adjustment for age and sex. Furthermore, an exploratory analysis was proBNP quartile are shown in Table 1. DTC patients in the highest quar-
performed with adjustment for age, sex, hypercholesterolemia, hyper- tiles of NT-proBNP were signicantly older, more often female, had a
tension, DM, BMI, and eGFR together (model 9). lower eGFR, and were more frequently diagnosed with hypertension,
The incremental value of NT-proBNP for cardiovascular events hypercholesterolemia, DM and pre-existent cardiovascular disease. At
and mortality risk prediction in DTC patients was analyzed using baseline, patients were in clinical follow-up for DTC for median [IQR]
Harrell's C-statistic. To this end the change in C-statistic attribut- 10.4 [4.118.5] years.
able to NT-proBNP was determined in an age- and sex-adjusted Median NT-proBNP level was 70 [40119] ng/L for patients with
model. Furthermore, we repeated crude, age-adjusted, and age- DTC, and 49 [2589] ng/L for controls (p b 0.001); see Fig. 2 for the
and sex-adjusted Cox regression analyses in subgroups of DTC patients, NT-proBNP distribution in both cohorts. Sixty-two DTC patients (23%)
who were diagnosed with DTC b 5 years, 5 to 15 years, or N15 years ago had an NT-proBNP 125 ng/L, as compared to 123 (15%) of control
at baseline. subjects.

Fig. 1. Flow chart of the study.

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
4 E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx

Table 1
Patient characteristics at the time of NT-proBNP measurement.

Parameter All patients Quartiles of NT-proBNP within DTC cohort p-Value

Quartile (NT-proBNP ng/L) 1 (540) 2 (4070) 3 (70119) 4 (1196809)

N 266 67 67 66 66
Mean age (years) SD 54.8 14.5 45.9 11.2 52.3 12.4 55.5 13.0 65.8 13.9 b0.001
Sex n (%) 0.001
Male 70 (26.3) 29 (43.3) 17 (25.4) 9 (13.6) 15 (22.7)
Female 196 (73.7) 38 (56.7) 50 (74.6) 57 (86.4) 51 (77.3)
TNM Tumor stage n (%)
T stage 0.297
Tx-T3 228 (85.7) 58 (86.6) 59 (88.1) 59 (89.4) 52 (78.8)
T4 38 (14.3) 9 (13.4) 8 (11.9) 7 (10.6) 14 (21.2)
N stage 0.173
Nx-N0 174 (65.4) 38 (56.7) 42 (62.7) 45 (68.2) 49 (74.2)
N1 92 (34.6) 29 (43.3) 25 (37.3) 21 (31.8) 17 (25.8)
M stage 0.203
Mx-M0 248 (93.2) 66 (98.5) 62 (92.5) 61 (92.4) 59 (89.4)
M1 18 (6.8) 1 (1.5) 5 (7.5) 5 (7.6) 7 (10.6)
Histology n (%) 0.167
Papillary 197 (74.1) 52 (77.6) 51 (76.1) 52 (78.8) 42 (63.6)
Follicular including Hrthle 69 (25.9) 15 (22.4) 16 (23.9) 14 (21.2) 24 (36.4)
Smoking n (%) 0.652
Present 45 (16.9) 13 (19.4) 11 (16.4) 14 (21.2) 7 (10.6)
Past 53 (19.9) 11 (16.4) 14 (20.9) 15 (22.7) 13 (19.7)
No 168 (63.2) 43 (64.2) 42 (62.7) 37 (56.1) 46 (69.7)
Body Mass Indexa Mean SD (kg/m2) 27.0 5.0 27.5 4.4 27.0 5.6 26.9 5.3 26.7 4.6 0.835
Hypertension n (%) 81 (30.5) 4 (6.0) 19 (28.4) 16 (24.2) 42 (63.6) b0.001
Hypercholesterolemia n (%) 34 (12.8) 3 (4.5) 12 (17.9) 3 (4.5) 16 (24.2) b0.001
Diabetes mellitus n (%) 16 (6.0) 0 (0) 6 (9.0) 2 (3.0) 8 (12.1) 0.010
eGFR (ml/min/1,73 m2)b Mean SD 82.1 20.4 89.4 22.4 82.3 16.6 82.1 19.0 74.4 20.7 b0.001
History of cardiovascular diseasec n (%) 14 (5.3) 0 (0) 0 (0) 1 (1.5) 13 (19.7) b0.001
Time in years from DTC diagnosis, median (IQR) 10.4 (4.118.5) 7.7 (3.713.9) 10.7 (3.817.0) 11.0 (4.419.5) 12.9 (6.621.8) 0.039

DTC = differentiated thyroid carcinoma.


TNM tumor stage and histology were determined at the time of DTC diagnosis.
a
16 missing BMI values were imputed.
b
7 missing eGFR values were imputed.
c
Consisted of a cardiac event in quartile 3; in quartile 4, 7 patients had a cardiac event, 2 heart failure and 4 a cerebrovascular event.

2.2. NT-proBNP as a prognostic marker in DTC patients respectively) had a cardiovascular event, and 38 patients (14.4%, n =
2, 5, 11 and 20 in quartiles 1, 2, 3, and 4, respectively) died (Table 2).
Median follow-up time for 263 DTC patients included for nal anal- Of DTC patients with an NT-proBNP 125 ng/L, two patients had clinical
ysis of NT-proBNP as a prognostic marker was 8.6 [IQR 6.69.0] years. A heart failure at baseline and nine developed this condition during fol-
total of 42 patients were lost to follow-up, median 4.5 [IQR 2.26.7] low-up.
years after NT-proBNP measurement (Fig. 1). During follow-up, 30 pa-
tients (11.4%, n = 2, 2, 4 and 22 in NT-proBNP quartiles 1, 2, 3 and 4,

Table 2
Outcome.

Variable DTC patients (n = 263)

Follow-up (years)
Median [IQR] 8.6 [6.69.0]
Time to cardiovascular event
Median [IQR] 4.0 [1.56.3]
Cardiovascular events n (%)a 30 (11.4)
New onset heart failure 10 (3.8)
Cerebrovascular events 8 (3.0)
Cerebrovascular accident 5
Transient ischemic attack 3
Cardiac events 10 (3.8)
Acute coronary syndromes 7
PCI/CABG 3
Abdominal aortic aneurysm 1 (0.4)
Peripheral artery disease 1 (0.4)
Mortality n (%)
All-cause mortality 38 (14.4)
Cardiovascular mortality 7 (2.7)
DTC mortality 11 (4.2)
Other death causes 19 (7.2)
Unknown cause of death 1 (0.4)

Fig. 2. Distribution of NT-proBNP levels in patients with DTC (orange bars) and controls DTC = differentiated thyroid carcinoma, PCI = percutaneous coronary intervention,
(blue bars). (For interpretation of the references to color in this gure legend, the reader CABG = coronary artery bypass graft.
a
is referred to the web version of this article.) Only the rst event per patient after NT-proBNP measurement is shown.

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx 5

this model, the C-index increased to 0.88 (95% CI 0.810.95), which


was an improvement of 0.064 (0.0030.126), p = 0.039. For all-cause
mortality, addition of NT-proBNP to age and sex did not signicantly im-
prove the C-statistic (data not shown).
The predictive value of NT-proBNP among patients diagnosed with
DTC b 5, 5 to 15, or N15 years ago is shown in Table 4. For all of these cat-
egories of DTC patients, a higher NT-proBNP was signicantly associated
with cardiovascular events in crude, age-adjusted, and age- and sex-ad-
justed analysis. NT-proBNP was associated as well with all-cause mor-
tality in crude analysis among all patient categories, but after
adjustment signicance was lost in some patient categories, although
HRs remained well above 1.

3. Discussion

In the present study we show that NT-proBNP is increased in DTC


patients. Furthermore, high NT-proBNP levels are independently associ-
ated with an increased risk for cardiovascular events and all-cause mor-
tality in patients with DTC.
During follow-up, DTC patients are in an exogenous (subclinical) hy-
perthyroid state for a considerable time consequent to THST. In line
with a previous nding in hyperthyroidism [14], we showed that NT-
proBNP levels are N 40% higher in DTC patients as compared to age-
and sex-matched controls. When evaluating alterations in NT-proBNP
in specic populations, use of age- and sex-matched controls is vital
since both increasing age and female sex are associated with higher
NT-proBNP levels [23]. This applies to DTC patients as well, as approxi-
mately 75% of DTC patients is female. Of further note, 62 DTC patients
(23%, of which only two had been clinically diagnosed with heart failure
at baseline) had an NT-proBNP level 125 ng/L, which has been pro-
posed the optimum cut-off point for excluding heart failure in a non-
acute setting [24]. Although the high NT-proBNP levels in DTC patients
can be explained by a the (subclinical) hyperthyroid state during at
least a part of follow-up, it can also be argued that DTC patients with
an NT-proBNP level 125 ng/L should be evaluated for undiagnosed
or subclinical heart failure. Further research is needed to elucidate the
meaning of high NT-proBNP levels in patients with DTC.
Earlier studies indicated a potential value of NT-proBNP as biomark-
er for cardiovascular events and all-cause mortality in cancer survivors
other than DTC [2527]. Here we showed, to our knowledge for the
rst time, that NT-proBNP may serve as a prognostic marker of in-
creased cardiovascular risk in DTC patients. In line, addition of NT-
Fig. 3. Kaplan-Meier survival curves of cardiovascular events (a) and all-cause mortality proBNP signicantly improved risk prediction (based on age and sex)
(b) for patients with DTC, by quartiles of NT-proBNP. for cardiovascular events in DTC patients. This indicates that the possi-
ble interaction between thyroid hormones and NT-proBNP does not im-
The Kaplan-Meier curves show that occurrence of both endpoints in- pede its value as a prognostic marker for cardiovascular risk. Although
creased according to increasing NT-proBNP quartiles (log-rank p b 0.001 there is a continuous scale (the higher NT-proBNP levels, the higher
for both, Fig. 3A and B). the risk), DTC patients in quartile 4 (NT-proBNP higher than 119 ng/L)
The results of the Cox regression models for DTC patients are shown had the worst prognosis for cardiovascular events and all-cause mortal-
in Table 3. In crude analyses (model 1), each SD increase in log NT- ity. When DTC patients have such high NT-proBNP values, evaluation
proBNP was associated with an HR (95% CI) of 4.07 (2.855.80) and and treatment of cardiovascular risk factors might be considered, as
2.64 (2.033.44) for cardiovascular and all-cause mortality, respective- well as evaluation of subclinical cardiovascular damage as cardiac sys-
ly. After adjustment for age and sex (model 2), HRs were 3.22 (2.17 tolic or diastolic dysfunction or left ventricular hypertrophy. Further-
4.79) for cardiovascular events and 1.61 (1.172.23) for all-cause mor- more, the degree of thyroid hormone suppression therapy could be
tality. In further Cox-regression models with adjustment for age and re-evaluated.
sex, and one of the covariates hypercholesterolemia, hypertension, Strengths of this study include the prospectively identied large co-
DM, BMI, eGFR, and tumor histology/DTC risk classication, NT-proBNP hort of DTC patients, and the duration of follow-up after NT-proBNP
remained signicantly associated with outcome. In the exploratory measurement of median 8.6 years. Moreover, to our knowledge this is
analysis with full adjustment (for age, sex, hypercholesterolemia, hy- the rst study in which NT-proBNP is compared between DTC patients
pertension, DM, BMI, and eGFR), NT-proBNP remained associated with and age- and sex-matched controls, and in which NT-proBNP is assessed
cardiovascular events and all-cause mortality, HRs (95% CIs) 3.38 as a prognostic marker for cardiovascular risk and mortality in patients
(2.145.35) and 1.83 (1.232.72), respectively. with DTC.
The Harrell's C-index of the Cox regression model including age and An important limitation of this study is that clinical and laboratory
sex for cardiovascular events was 0.82 (95% CI 0.760.88), indicating measurements including NT-proBNP levels were determined at least
that 82% of patients with DTC were correctly classied with regard to 1 year after DTC diagnosis, but otherwise irrespective of follow-up
development of a cardiovascular event. After addition of NT-proBNP to after initial DTC diagnosis and treatment, rather than on a standardized

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
6 E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx

Table 3
Cox regression models for cardiovascular events and all-cause mortality.

DTC patients (n = 263) Cardiovascular events All-cause mortality

HR (95% CI) p-Value HR (95% CI) p-Value

Model 1 (crude)
Log NT-proBNP per SD increase 4.07 (2.855.80) b0.001 2.64 (2.033.44) b0.001
Model 2
Log NT-proBNP per SD increase 3.22 (2.174.79) b0.001 1.61 (1.172.23) 0.004
Model 3
Log NT-proBNP per SD increase 3.18 (2.134.74) b0.001 1.59 (1.122.24) 0.009
Model 4
Log NT-proBNP per SD increase 3.08 (2.034.67) b0.001 1.57 (1.112.21) 0.011
Model 5
Log NT-proBNP per SD increase 3.23 (2.184.79) b0.001 1.82 (1.282.59) 0.001
Model 6
Log NT-proBNP per SD increase 3.35 (2.175.18) b0.001 1.61 (1.172.23) 0.004
Model 7
Log NT-proBNP per SD increase 3.23 (2.144.85) b0.001 1.89 (1.322.69) b0.001
Model 8
Log NT-proBNP per SD increase 3.75 (2.405.86) b0.001 1.54 (1.082.18) 0.016
Model 9
Log NT-proBNP per SD increase 3.38 (2.145.35) b0.001 1.83 (1.232.72) 0.003

SD = standard deviation, log NT-proBNP SD was 0.45.


Model 2: Adjusted for age and sex.
Model 3: Adjusted for age, sex, and hypercholesterolemia.
Model 4: Adjusted for age, sex, and hypertension.
Model 5: adjusted for age, sex, and diabetes mellitus (DM).
Model 6: adjusted for age, sex, and BMI.
Model 7: adjusted for age, sex, and eGFR.
Model 8: adjusted for age, sex, tumor histology and DTC risk classication.
Model 9: adjusted for age, sex, hypercholesterolemia, hypertension, DM, BMI, and eGFR.

time-point. Therefore, we explored whether the association between Also, the possibility of residual confounding due to unmeasured con-
NT-proBNP and outcome is affected by the duration of DTC before founders and some errors in measurements of potential confounders
entry in the study. Although these analyses were hampered by low sta- cannot be excluded. We did not compare differences in cardiovascular
tistical power, we could indicate that (at least in crude analyses) NT- event rates between the two cohorts, due to differences in end-point as-
proBNP is associated with cardiovascular events and all-cause mortality certainment. Furthermore, we did not evaluate whether outcome was
independent from the duration of DTC. Another limitation is that the associated with TSH levels, since the current study design was not ap-
DTC study population may be biased by survival, as only patients who propriate to do so. Finally, it would have been interesting to correlate
survived and were in follow-up until the period of 20042008 were el- triiodothyronine (T3) and thyroxine (T4) values with NT-proBNP, but
igible for inclusion. Patients with aggressive disease (who are more like- unfortunately these were not available.
ly to die during follow-up) may therefore be underrepresented in this We conclude that NT-proBNP levels are elevated in DTC patients,
study. Still, we consider our population suitable for the aim of the cur- and that higher NT-proBNP levels are independently associated with
rent study, in view of the clinical relevance of NT-proBNP representing an increased risk for cardiovascular events and all-cause mortality in pa-
an adequate marker of increased cardiovascular risk in DTC patients tients with DTC. Determination of NT-proBNP may contribute to an im-
with a considerable life expectancy. Furthermore, only combined end- proved care for DTC patients, as it may reveal patients at increased risk
points of all cardiovascular events and all-cause mortality were ana- for cardiovascular events and mortality. A more stringent approach to
lyzed, since separate analysis of their components lacked statistical cardiovascular risk management in these patients with identication
power. Besides, we had limited power in the multivariate analyses for and treatment of cardiovascular risk factors may prevent cardiovascular
adjustment of multiple cardiovascular risk factors at the same time. events, and eventually improve survival and quality of life.

Table 4
Subgroup analysis within DTC cohort of DTC patients diagnosed b5, 5 to 15, and N15 years before entry in the study.

Subgroup analysis DTC patients Cardiovascular events All-cause mortality

Years since DTC diagnosis at baseline HR (95% CI) p-Value HR (95% CI) p-Value

b5 years No of events 5 15
n = 75 patients Crude 14.9 (3.1870.0) 0.001 4.01 (2.117.64) b.001
Age-adjusted 7.85 (1.4942.6) 0.017 1.57 (0.693.61) 0.286
Age- and sex adjusted 8.23 (1.3948.8) 0.020 1.57 (0.673.65) 0.297
5 to 15 years No of events 11 13
n = 101 patients Crude 3.09 (1.625.87) 0.001 2.77 (1.525.03) 0.001
Age-adjusted 2.06 (1.014.20) 0.047 1.95 (1.013.77) 0.048
Age- and sex adjusted 2.05 (1.014.16) 0.046 1.91 (0.983.74) 0.058
N15 years No of events 14 10
n = 87 patients Crude 3.53 (2.135.86) b0.001 2.82 (1.824.38) b0.001
Age-adjusted 3.64 (1.976.71) b0.001 1.85 (1.073.20) 0.028
Age- and sex adjusted 4.27 (2.347.78) b0.001 2.03 (1.163.57) 0.014

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020
E.N. Klein Hesselink et al. / Clinical Biochemistry xxx (2017) xxxxxx 7

Declaration of interest [13] G.C. Linssen, S.J. Bakker, A.A. Voors, et al., N-terminal pro-B-type natriuretic peptide
is an independent predictor of cardiovascular morbidity and mortality in the general
population, Eur. Heart J. 31 (1) (2010) 120127.
We declare that there is no conict of interest that could be per- [14] K. Kato, H. Murakami, O. Isozaki, T. Tsushima, K. Takano, Serum concentrations of
ceived as prejudicing the impartiality of the research reported. BNP and ANP in patients with thyrotoxicosis, Endocr. J. 56 (1) (2009) 1727.
[15] F. Liang, P. Webb, A. Marimuthu, S. Zhang, D.G. Gardner, Triiodothyronine increases
brain natriuretic peptide (BNP) gene transcription and amplies endothelin-depen-
Funding dent BNP gene transcription and hypertrophy in neonatal rat ventricular myocytes,
J. Biol. Chem. 278 (17) (2003) 1507315083.
[16] H.J. Lambers Heerspink, A.H. Brantsma, D. de Zeeuw, et al., Albuminuria assessed
Not applicable. from rst-morning-void urine samples versus 24-hour urine collections as a predic-
tor of cardiovascular morbidity and mortality, Am. J. Epidemiol. 168 (8) (2008)
References 897905.
[17] D. van Dijk, J.T. Plukker, A.N. van der Horst-Schrivers, et al., The value of detectable
[1] F.A. Verburg, U. Mader, K. Tanase, et al., Life expectancy is reduced in differentiated thyroglobulin in patients with differentiated thyroid cancer after initial 131I therapy,
thyroid cancer patients 45 years old with extensive local tumor invasion, lateral Clin. Endocrinol. 74 (1) (2011) 104110.
lymph node, or distant metastases at diagnosis and normal in all other DTC patients, [18] K.M. van Tol, P.L. Jager, E.G. de Vries, et al., Outcome in patients with differentiated
J. Clin. Endocrinol. Metab. 98 (1) (2013) 172180. thyroid cancer with negative diagnostic whole-body scanning and detectable stim-
[2] E.N. Klein Hesselink, J.D. Lefrandt, E.P. Schuurmans, et al., Increased risk of atrial - ulated thyroglobulin, Eur. J. Endocrinol. 148 (6) (2003) 589596.
brillation after treatment for differentiated thyroid carcinoma, J. Clin. Endocrinol. [19] Comprehensive Cancer Centre the Netherlands, Oncoline, Cancer Clinical Practice
Metab. 100 (12) (2015) 45634569. Guidelines, http://oncoline.nl/index.php?language=en Accessed September 28,
[3] R.M. Abdulrahman, V. Delgado, H.C. Hoftijzer, et al., Both exogenous subclinical hy- 2012.
perthyroidism and short-term overt hypothyroidism affect myocardial strain in pa- [20] J.D. Brierley, T. Panzarella, R.W. Tsang, M.K. Gospodarowicz, B. O'Sullivan, A compar-
tients with differentiated thyroid carcinoma, Thyroid 21 (5) (2011) 471476. ison of different staging systems predictability of patient outcome. Thyroid carcino-
[4] J.W. Smit, C.F. Eustatia-Rutten, E.P. Corssmit, et al., Reversible diastolic dysfunction ma as an example, Cancer 79 (12) (1997) 24142423.
after long-term exogenous subclinical hyperthyroidism: a randomized, placebo- [21] P.O. Collinson, S.C. Barnes, D.C. Gaze, G. Galasko, A. Lahiri, R. Senior, Analytical per-
controlled study, J. Clin. Endocrinol. Metab. 90 (11) (2005) 60416047. formance of the N terminal pro B type natriuretic peptide (NT-proBNP) assay on
[5] M.K. Horne 3rd, K.K. Singh, K.G. Rosenfeld, et al., Is thyroid hormone suppression the elecsys 1010 and 2010 analysers, Eur. J. Heart Fail. 6 (3) (2004) 365368.
therapy prothrombotic? J. Clin. Endocrinol. Metab. 89 (9) (2004) 44694473. [22] J. Carlin, N. Li, P. Greenwood, C. Coffey, Clinical epidemiology and biostatistics unit.
[6] E.N. Klein Hesselink, M.S. Klein Hesselink, G.H. de Bock, et al., Long-term cardiovas- Tools for analyzing multiple imputed datasets, Stata J. 3 (3) (2003) 226244.
cular mortality in patients with differentiated thyroid carcinoma: an observational [23] A. Luchner, G. Behrens, J. Stritzke, et al., Long-term pattern of brain natriuretic pep-
study, J. Clin. Oncol. 31 (32) (2013) 40464053. tide and N-terminal pro brain natriuretic peptide and its determinants in the gener-
[7] L.B. Daniels, A.S. Maisel, Natriuretic peptides, J. Am. Coll. Cardiol. 50 (25) (2007) al population: contribution of age, gender, and cardiac and extra-cardiac factors, Eur.
23572368. J. Heart Fail. 15 (8) (2013) 859867.
[8] A. Clerico, A. Giannoni, S. Vittorini, C. Passino, Thirty years of the heart as an endo- [24] J.J. McMurray, S. Adamopoulos, S.D. Anker, et al., ESC guidelines for the diagnosis
crine organ: physiological role and clinical utility of cardiac natriuretic hormones, and treatment of acute and chronic heart failure 2012: the task force for the diagno-
Am. J. Physiol. Heart Circ. Physiol. 301 (1) (2011) H12H20. sis and treatment of acute and chronic heart failure 2012 of the european society of
[9] E.R. Levin, D.G. Gardner, W.K. Samson, Natriuretic peptides, N. Engl. J. Med. 339 (5) cardiology. Developed in collaboration with the heart failure association (HFA) of
(1998) 321328. the ESC, Eur. J. Heart Fail. 14 (8) (2012) 803869.
[10] R.S. Gardner, F. Ozalp, A.J. Murday, S.D. Robb, T.A. McDonagh, N-terminal pro-brain [25] C.M. Korse, B.G. Taal, C.A. de Groot, R.H. Bakker, J.M. Bonfrer, Chromogranin-A and
natriuretic peptide. A new gold standard in predicting mortality in patients with ad- N-terminal pro-brain natriuretic peptide: an excellent pair of biomarkers for diag-
vanced heart failure, Eur. Heart J. 24 (19) (2003) 17351743. nostics in patients with neuroendocrine tumor, J. Clin. Oncol. 27 (26) (2009)
[11] T. Omland, A. Persson, L. Ng, et al., N-terminal pro-B-type natriuretic peptide and 42934299.
long-term mortality in acute coronary syndromes, Circulation 106 (23) (2002) [26] S.E. Lipshultz, D.C. Landy, G. Lopez-Mitnik, et al., Cardiovascular status of childhood
29132918. cancer survivors exposed and unexposed to cardiotoxic therapy, J. Clin. Oncol. 30
[12] C. Kistorp, I. Raymond, F. Pedersen, F. Gustafsson, J. Faber, P. Hildebrandt, N-terminal (10) (2012) 10501057.
pro-brain natriuretic peptide, C-reactive protein, and urinary albumin levels as pre- [27] C.A. Brouwer, A. Postma, J.M. Vonk, et al., Systolic and diastolic dysfunction in long-
dictors of mortality and cardiovascular events in older adults, JAMA 293 (13) (2005) term adult survivors of childhood cancer, Eur. J. Cancer 47 (16) (2011) 24532462.
16091616.

Please cite this article as: E.N. Klein Hesselink, et al., NT-proBNP is increased in differentiated thyroid carcinoma patients and may predict
cardiovascular risk, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.02.020

You might also like