CLINICAL CARDIOLOGY

Cardiology Journal
2018, Vol. 25, No. 3, 371–376
DOI: 10.5603/CJ.a2017.0075
Copyright © 2018 Via Medica
ORIGINAL ARTICLE ISSN 1897–5593

Blood urea nitrogen in the prediction of in-hospital

mortality of patients with acute aortic dissection
Jun Liu 1, Lian-Lian Sun1, Jue Wang 2, Guang Ji 1
1
Division of Cardiology, the First Affiliated Hospital of Wenzhou Medical University,
Wenzhou, Zhejiang, P. R. China
2
Division of Cardiothoracic Surgery, the First Affiliated Hospital of
Wenzhou Medical University, Wenzhou, Zhejiang, P. R. China

Abstract
Background: Blood urea nitrogen (BUN) has been shown to be associated with adverse cardiovascular
disease outcomes. The aim of the present study was to evaluate the prognostic role of BUN in patients
with acute aortic dissection (AAD). Hypothesis: BUN has correlation with in-hospital mortality of
patients with AAD.
Methods: Patients admitted to the emergency room within the first 24 h of onset of AAD were included
in the study. BUN levels were measured on admission and the endpoints were mortality during hospi-
talization after receiving surgical or endovascular repair.
Results: A total of 192 patients with AAD were enrolled. During hospitalization, 19 patients died and
173 patients survived. Increased levels of BUN (8.9 [7.0–9.7] vs. 6.0 [5.1–7.2] mmol/L, p < 0.001)
were found in non-survivors compared with those survived. Using multivariable logistic analysis,
BUN was an independent predictor of in-hospital mortality in patients with AAD (OR 1.415, 95% CI
1.016–1.971, p = 0.040). Furthermore, using receiver operating characteristic analysis, the optimal
cutoff value for BUN was 6.95 mmol/L. Under this value, the area under the curve was 0.785 (95% CI
0.662–0.909, p < 0.001) and the sensitivity and specificity to predict in-hospital mortality was 78.9%,
and 72.2%, respectively.
Conclusions: Admission BUN levels were an independent predictor for in hospital mortality in pa-
tients with AAD. (Cardiol J 2018; 25, 3: 371–376)
Key words: blood urea nitrogen, acute aortic dissection, in-hospital mortality

Introduction in predicting poor clinical outcomes [1, 3, 4].

In fact, BUN has also been demonstrated to be
The relationship between renal dysfunction a valid biomarker for disease severity and progno-
and adverse cardiovascular disease outcomes sis in many other conditions such as community-
have been well established in patients with coro- acquired pneumonia [5], acute pancreatitis [6] and
nary artery disease (CAD) [1] and in those with acute intracerebral hemorrhage [7].
heart failure [2]. Prior studies utilizing serum Acute aortic dissection (AAD) is a life-threat-
creatinine, estimates of glomerular filtration ening disorder that necessitates immediate man-
rate (eGFR) or blood urea nitrogen (BUN) to agement. Previous studies have demonstrated that
investigate the prognostic role of renal dysfunc- kidney injury is a risk factor for both short and
tion in cardiovascular disease have found that long term mortality in patients with AAD [8, 9].
BUN is more sensitive than creatinine and eGFR However, these studies used serum creatinine

Address for correspondence: Jun Liu, MD, Division of Cardiology, the First Affiliated Hospital of Wenzhou Medical
University, Wenzhou, Zhejiang, P. R. China, tel: +86 577 55579233, e-mail: liujunemail1@163.com
Received: 08.03.2017 Accepted: 20.05.2017

www.cardiologyjournal.org 371
Cardiology Journal 2018, Vol. 25, No. 3

as a surrogate marker for renal dysfunction and Statistical analysis

whether BUN could also provide some prognostic Normality of continuous data was assessed
significance in these patients is unknown. Hence, using the Shapiro-Wilk test. Continuous variables
the aim of the present study was to evaluate the are presented as the mean ± standard deviation or
relationship between BUN and in hospital mortality median and interquartile range according to whether
in patients admitted with AAD. they follow normal distributions. Categorical vari-
ables are presented as proportions. Comparisons be-
Methods tween groups were performed with unpaired Student
t tests for normally distributed continuous variables
Study population and Wilcoxon Mann-Whitney tests for non-normally
Herein is retrospectively investigated the distributed continuous variables. Categorical vari-
medical records of patients with AAD admitted to ables were compared by c2 tests or Fisher’s exact test
the First Affiliated Hospital of Wenzhou Medical when appropriate. Univariate analysis and multiple
University between December 2012 and June 2016. logistic regression analysis were used to identify the
The study was reviewed and approved by the ethics predictors of in-hospital mortality. Receiver operat-
committee of the hospital and informed consent ing characteristic (ROC) analysis was performed to
was waived due to its retrospective nature. determine the cut-off value for BUN in predicting in-
The diagnosis of AAD was confirmed by multi- hospital mortality with high sensitivity and specificity.
detector computed tomography scan and the type A p value < 0.05 was considered to be statistically sig-
of AAD was classified in accordance with Stanford nificant. All the statistical analyses were performed
University criteria. Patients were included in the using SPSS statistical software, version 19.0 (SPSS
present study if they fulfilled the following criteria: Inc., Chicago, Illinois).
(1) the time interval between symptoms onset
and hospital admission of ≤ 24 h; (2) receiving Results
surgical or endovascular repair for AAD du­
ring hospitalization. Exclusion criteria included: Baseline clinical characteristics
(1) presence of AAD for more than 24 h; (2) diagnosis Six hundred twenty six patients were identi-
with Marfan syndrome; (3) prior history of aortic fied with a diagnosis of AD, of whom 251 were
dissection (AD). admitted within 24 h after onset of symptoms.
Among them, 8 patients diagnosed with Marfan
Treatment syndrome, 9 patients with prior history of AD and
For patients with confirmed diagnosis of AD, 42 patients died during hospitalization with con-
urapidil, sodium nitroprusside or nitroglycerine servative treatment were excluded from analysis.
were was administered intravenously to reduce A total of 192 patients met our eligibility criteria
systolic blood pressure (SBP) to 100–120 mmHg. and were included in the present study. During
Beta-blocker was administered to all patients ex- hospitalization, 19 patients died and 173 patients
cept those with the contraindication. survived.
All type A AD patients and a small part of Patient characteristics are shown in Table 1.
type B AD patients underwent surgery repair via There were no significant differences in sex,
cardiopulmonary bypass. Endovascular repair was hypertension, CAD, dyslipidemia, DM, smoking,
performed using commercially available endografts alcohol use, aortic diameter, heart rate and neu-
with the patients under general anesthesia. trophil counts between the two groups. Compared
with survivors, the non-survivors were older
Data collection (62 [50–66] vs. 50 [44–60] years, p = 0.03), had
The following clinical data were collected on a higher percentage of type A AD (78.9% vs.
admission: age, sex, presence of hypertension, CAD, 43.9%, p = 0.003). The non-survivors had a sig-
dyslipidemia, diabetes mellitus (DM), smoking and nificantly lower SBP (116.2 ± 26.0 vs. 145.6 ±
drinking habits, type of AAD, aortic diameter, blood ± 26.0 mmHg, p < 0.001), diastolic blood pres-
pressure, heart rate, and laboratory data. sure (DBP; 66.6 ± 15.5 vs. 80.2 ± 18.6 mmHg,
p = 0.003), and hemoglobin (122 [111–133] vs.
Endpoint 132 [122–144] g/L, p = 0.006) on admission. In
The study endpoint was defined as all-cause addition, white blood cell counts (14.8 [13.1–18.0]
mortality during hospitalization. vs. 13.2 [11.2–15.8] ×109/L, p = 0.047), creatinine

372 www.cardiologyjournal.org
 Jun Liu et al., Blood urea nitrogen and in-hospital mortality of acute aortic dissection

Table 1. Baseline characteristics of the patients.

Variable All patients (n = 192) Survivor (n = 173) Non-survivor (n = 19) P

Age [years] 51 (44-62) 50 (44-60) 62 (50-66) 0.03
Male 151 (78.6%) 135 (78.0%) 16 (84.2%) 0.388
Hypertension 137 (71.4%) 122 (70.5%) 15 (78.9%) 0.316
Coronary artery disease 1 (0.5%) 1 (0.6%) 0 (0%) 0.901
Dyslipidemia 1 (0.5%) 1 (0.6%) 0 (0%) 0.901
Diabetes mellitus 7 (3.6%) 6 (3.5%) 1 (5.3%) 0.524
Smoking 66 (34.4%) 61 (35.3%) 5 (26.3%) 0.612
Alcohol use 41 (21.4%) 39 (22.5%) 2 (10.5%) 0.181
Type A AD 91 (47.4%) 76 (43.9%) 15 (78.9%) 0.003
SBP [mmHg] 142.7 ± 27.4 145.6±26.0 116.2 ± 26.0 < 0.001
DBP [mmHg] 78.8 ± 18.7 80.2 ± 18.6 66.6 ± 15.5 0.003
Aortic diameter [mm] 39 (36–43) 39 (36–43) 41 (35–45) 0.329
Heart rate [bpm] 80.9 ± 16.6 80.5±16.3 84.4 ± 19.0 0.333
White blood cell [×109/L] 13.5 (11.2–16.0) 13.2 (11.2–15.8) 14.8 (13.1–18.0) 0.047
Neutrophil [×10 /L]9
12.0 ± 3.7 11.8 ± 3.6 13.3 ± 4.3 0.100
Hemoglobin [g/L] 132 (122–144) 132 (122–144) 122 (111–133) 0.006
Creatinine [mmol/L] 78.0 (64.0–102.8) 76.0 (63.0–96.0) 127.0 (83.0–150.0) < 0.001
BUN [mmol/L] 6.1 (5.2–7.6) 6.0 (5.1–7.2) 8.9 (7.0–9.7) < 0.001

Data are expressed as mean ± standard deviation, median (25th–75th percentile) or the number (percentages) of patients. The bold values
indicate statistical significance; AD — aortic dissection; BUN — blood urea nitrogen; DBP — diastolic blood pressure; SBP — systolic blood
pressure

Table 2. Univariable logistic regression for in-hospital mortality.

Variable OR 95% CI P
Age 1.037 0.999–1.075 0.0054
Type A AD 4.786 1.526–15.011 0.007
SBP [mmHg] 0.957 0.929–0.976 < 0.001
DBP [mmHg] 0.960 0.934–0.987 0.004
White blood cell [×10 /L] 9
1.116 0.997–1.249 0.057
Hemoglobin [g/L] 0.965 0.940–0.992 0.010
Creatinine [mmol/L] 1.005 1.000–1.010 0.037
BUN [mmol/L] 1.330 1.118–1.582 0.001

CI — confidence interval; OR — odds ratio; rest abbreviations as in Table 1

(127.0 [83.0–150.0] vs. 76.0 [63.0–96.0] mmol/L, white blood cell counts, hemoglobin, creatinine
p < 0.001) and BUN (8.9 [7.0–9.7] vs. 6.0 [5.1–7.2] and BUN. Except for white blood cell counts, all
mmol/L, p < 0.001) levels were higher in the non- variables included were significantly associated
survivor group than in the survivor group. with in-hospital mortality in univariable logistic
regression (Table 2). The multivariable logistic
Predictors for in-hospital mortality regression model for in-hospital mortality included
Logistic regression was performed to iden- all variables from the univariable analysis. The only
tify potential predictors for in-hospital mortality variables that remained as independent predictors
in patients with AAD. Variables included in the of in-hospital mortality were age (odds ratio [OR]
univariable logistic regression analysis for in- 1.059, 95% confidence interval [CI] 1.003–1.119,
hospital mortality were age, type A AD, SBP, DBP, p = 0.039), SBP (OR 0.962, 95% CI 0.928–0.998,

www.cardiologyjournal.org 373
Cardiology Journal 2018, Vol. 25, No. 3

Table 3. Multivariable logistic regression for in-hospital mortality.

Variable OR 95% CI P
Age 1.059 1.003–1.119 0.039
Type A AD 2.099 0.451–9.774 0.345
SBP [mmHg] 0.962 0.928–0.998 0.037
DBP [mmHg] 1.000 0.955–1.047 0.992
White blood cell [×109/L] 1.113 0.946–1.310 0.196
Hemoglobin [g/L] 0.983 0.943–1.026 0.435
Creatinine [mmol/L] 0.996 0.984–1.007 0.469
BUN [mmol/L] 1.415 1.016–1.971 0.040

CI — confidence interval; OR — odds ratio; rest abbreviations as in Table 1.

Figure 1. Odds ratios of blood urea nitrogen for predicting in-hospital mortality according to the type of acute aortic
dissection (AAD) and treatment; CI — confidence interval; EVR — endovascular repair.

Table 4. Diagnostic value of blood urea nitrogen for in-hospital mortality.

AUC Cut-off value SE 95% CI P Sensitivity Specificity

0.785 6.95 0.063 0.662–0.909 < 0.001 0.789 0.722

AUC — area under the curve; CI — confidence interval; SE — standard error

p = 0.0037) and BUN (OR 1.415, 95% CI 1.016– ROC analysis

–1.971, p = 0.040) (Table 3). ROC analysis yielded an area under the curve
Given the high mortality rates in patients (AUC) statistic of 0.785 (95% CI 0.662–0.909,
with type A AD and those who undergo surgery p < 0.001) and the cut-off value of BUN to predict
treatment, subgroup analysis was performed by in-hospital mortality was 6.95 (sensitivity 78.9%,
the type of AD and treatment. In patients with specificity 72.2%) (Table 4, Fig. 2).
type A AD, BUN was associated with high
risk of in-hospital mortality (OR 1.663, 95% Discussion
CI 1.006–2.747, p = 0.047). For patients with
type B AD, BUN was not an independent pre- The main findings of the present study were
dictor for in-hospital mortality. As for the sub- that admission BUN levels were independently as-
group, stratified by surgical or endovascular sociated with in-hospital mortality in patients with
repair, BUN independently predicted in-hospi- AAD admitted within 24 h after onset of symptoms.
tal mortality (OR 1.811, 95% CI 1.190–2.756, When BUN was ≥ 6.95 mmol/L, the sensitivity and
p = 0.006; OR 1.379, 95% CI 1.023–1.859, specificity for in-hospital mortality were 78.9% and
p = 0.035, respectively) (Fig. 1). 72.2%, respectively.

374 www.cardiologyjournal.org
 Jun Liu et al., Blood urea nitrogen and in-hospital mortality of acute aortic dissection

rapidly increase urea permeability in the collecting

duct through phosphorylation and apical plasma-
membrane accumulation of the urea transporter
A1 (UT-A1) and subsequently lead to increased
reabsorption of urea in the collecting duct [13].
Therefore, increased BUN level could be regarded
as a surrogate marker for hemodynamic and neu-
rohormonal alternations in AAD.
The prognostic implication of increased BUN
has been evaluated in patients with acute de-
compensated heart failure (ADHF). Studies have
demonstrated that in the setting of ADHF, patients
with high admission BUN level had low cumulative
survival rate [14, 15]. In addition, Miura et al. [16]
have shown that in patients admitted for ADHF, a
BUN increase during hospitalization also indicated
worse long-term prognosis, independent of renal
function. Moreover, increased BUN is also associ-
ated with worse prognosis in a wide spectrum of
Figure 2. Receiver operating characteristic curve of patients with acute coronary syndromes [1, 17].
blood urea nitrogen for predicting in-hospital mortality However, few data are available on the relationship
in patients with acute aortic dissection. between BUN and clinical outcomes in patients
with AAD. The present findings indicate for the
first time that in patients with AAD receiving
Urea is synthesized in liver via protein catabo- surgical or endovascular repair, high admission
lism and blood urea is freely filtered at the glo- BUN levels were associated with high in-hospital
merulus and undergoes tubular reabsorption. In the mortality. BUN may be a promising marker for risk
absence of enhanced protein catabolism, such as se- stratification in patients with AAD.
vere infection and burning, trauma, glucocorticoids
therapy or high protein diet, urea levels in blood is Limitations of the study
determined by the GFR and tubular reabsorption. This study has several limitations. First, be-
In the setting of AAD, hemodynamic disturbances cause of its observational nature, some factors that
lead to activation of the neurohormonal axis, re- may have influenced the outcome of AD were
sulting in increased sympathetic nervous system unavailable, such as D-dimer and C-reactive protein.
[10] and renin–angiotensin–aldosterone system Second, this study was carried out in patients receiv-
activity [11]. Increased angiotensin and adrenergic ing surgical or endovascular repair for AAD, exclud-
stimulation decrease renal urea excretion via ef- ing those receiving conservative treatment, which
fects on vascular, glomerular, and tubular effects on could limit results to extrapolate to a wider group
the kidney. These neurohormonal responses cause of patients. Furthermore, this study was carried
renal vasoconstriction, decreases in glomerular out in a single institution and enrolled a relatively
ultrafiltration, and increases in proximal tubular small number of patients, which precluded subgroup
sodium and water reabsoption. Consequently, analysis. A future prospective study with a larger
the fluid delivery in the collecting conduct will sample size is required to confirm these findings.
decrease and urine flow will slow. Because urea
reabsorption in the distal tubule is urine flow Conclusions
dependent, the decreased slowing of tubular flow
will enhance urea reabsorption [12]. Besides, AAD In summary, the present study found that
can cause low cardiac output by forming pericar- elevated BUN levels were an independent predic-
dial effusion or myocardial ischemia secondary to tor of in-hospital mortality in patients with AAD
coronary artery involvement. In order to preserve receiving surgical or endovascular repair. Serum
systemic perfusion, arginine vasopressin (AVP) is BUN may serve as a simple marker to identify high
released as a compensatory mechanism. AVP can risk patients.

www.cardiologyjournal.org 375
Cardiology Journal 2018, Vol. 25, No. 3

Funding: This study was supported by a grant Neurol Res. 2012; 34(5): 518–521, doi: 10.1179/1743132812Y.00
from the Natural Science Foundation of Zhejiang 00000041, indexed in Pubmed: 22664363.
8. Takahashi T, Hasegawa T, Hirata N, et al. Impact of acute kid-
Province (LQ17H020005).
ney injury on in-hospital outcomes in patients with DeBakey
type III acute aortic dissection. Am J Cardiol. 2014; 113(11):
Conflict of interest: None declared 1904–1910, doi: 10.1016/j.amjcard.2014.03.023, indexed in Pub-
med: 24837272.
References 9. Ko T, Higashitani M, Sato A, et al. Impact of Acute Kidney Injury
on Early to Long-Term Outcomes in Patients Who Underwent
1. Kirtane AJ, Leder DM, Waikar SS, et al. TIMI Study Group. Se- Surgery for Type A Acute Aortic Dissection. Am J Cardiol. 2015;
rum blood urea nitrogen as an independent marker of subsequent 116(3): 463–468, doi: 10.1016/j.amjcard.2015.04.043, indexed in
mortality among patients with acute coronary syndromes and Pubmed: 26026862.
normal to mildly reduced glomerular filtration rates. J Am Coll 10. Zhipeng Hu, Zhiwei W, Lilei Yu, et al. Sympathetic hyperactivity
Cardiol. 2005; 45(11): 1781–1786, doi: 10.1016/j.jacc.2005.02.068, and aortic sympathetic nerve sprouting in patients with tho-
indexed in Pubmed: 15936606. racic aortic dissection. Ann Vasc Surg. 2014; 28(5): 1243–1248,
2. Brisco MA, Coca SG, Chen J, et al. Blood urea nitrogen/creati- doi: 10.1016/j.avsg.2013.11.016, indexed in Pubmed: 24440187.
nine ratio identifies a high-risk but potentially reversible form of 11. Li Y, Hu J, Qian H, et al. Novel findings: Expression of an-
renal dysfunction in patients with decompensated heart failure. giotensin-converting enzyme and angiotensin-converting
Circ Heart Fail. 2013; 6(2): 233–239, doi: 10.1161/CIRCHEART- enzyme 2 in thoracic aortic dissection and aneurysm. J Re-
FAILURE.112.968230, indexed in Pubmed: 23325460. nin Angiotensin Aldosterone Syst. 2015; 16(4): 1130–1134,
3. Smith GL, Shlipak MG, Havranek EP, et al. Serum urea nitrogen, doi: 10.1177/1470320314549219, indexed in Pubmed: 25237166.
creatinine, and estimators of renal function: mortality in older 12. Schrier RW. Blood urea nitrogen and serum creatinine: not
patients with cardiovascular disease. Arch Intern Med. 2006; married in heart failure. Circ Heart Fail. 2008; 1(1): 2–5, doi:
166(10): 1134–1142, doi: 10.1001/archinte.166.10.1134, indexed 10.1161/CIRCHEARTFAILURE.108.770834, indexed in Pub-
in Pubmed: 16717177. med: 19808263.
4. Klein L, Massie BM, Leimberger JD, et al. OPTIME-CHF Inves- 13. Sands JM, Blount MA, Klein JD. Regulation of Renal Urea
tigators. Admission or changes in renal function during hospitali- Transport by Vasopressin. Trans Am Clin Climatol Assoc. 2011;
zation for worsening heart failure predict postdischarge survival: 122(122): 82–92.
results from the Outcomes of a Prospective Trial of Intravenous 14. Aronson D, Mittleman MA, Burger AJ. Elevated blood urea ni-
Milrinone for Exacerbations of Chronic Heart Failure (OPTIME- trogen level as a predictor of mortality in patients admitted for
CHF). Circ Heart Fail. 2008; 1(1): 25–33, doi: 10.1161/CIR- decompensated heart failure. Am J Med. 2004; 116(7): 466–473,
CHEARTFAILURE.107.746933, indexed in Pubmed: 19808267. doi: 10.1016/j.amjmed.2003.11.014, indexed in Pubmed: 15047036.
5. Ugajin M, Yamaki K, Iwamura N, et al. Blood urea nitrogen to se- 15. Fonarow GC, Adams KF, Abraham WT, et al. ADHERE Scien-
rum albumin ratio independently predicts mortality and severity tific Advisory Committee, Study Group, and Investigators. Risk
of community-acquired pneumonia. Int J Gen Med. 2012; 5: 583– stratification for in-hospital mortality in acutely decompensated
–589, doi: 10.2147/IJGM.S33628, indexed in Pubmed: 22866010. heart failure: classification and regression tree analysis. JAMA.
6. Koutroumpakis E, Wu BU, Bakker OJ, et al. Admission Hema- 2005; 293(5): 572–580, doi: 10.1001/jama.293.5.572, indexed in
tocrit and Rise in Blood Urea Nitrogen at 24 h Outperform other Pubmed: 15687312.
Laboratory Markers in Predicting Persistent Organ Failure and 16. Miura M, Sakata Y, Nochioka K, et al. Prognostic impact of blood
Pancreatic Necrosis in Acute Pancreatitis: A Post Hoc Analysis urea nitrogen changes during hospitalization in patients with
of Three Large Prospective Databases. Am J Gastroenterol. acute heart failure syndrome. Circ J. 2013; 77(5): 1221–1228,
2015; 110(12): 1707–1716, doi: 10.1038/ajg.2015.370, indexed in indexed in Pubmed: 23392088.
Pubmed: 26553208. 17. Aronson D, Hammerman H, Beyar R, et al. Serum blood urea ni-
7. Rhoney DH, Parker D, Millis SR, et al. Kidney dysfunction at the trogen and long-term mortality in acute ST-elevation myocardial
time of intracerebral hemorrhage is associated with increased infarction. Int J Cardiol. 2008; 127(3): 380–385, doi: 10.1016/j.
in-hospital mortality: a retrospective observational cohort study. ijcard.2007.05.013, indexed in Pubmed: 17765341.

376 www.cardiologyjournal.org