Methods

The ATTEND registry is a nationwide, multicenter, patient-based, prospective cohort study involving 52 medical hospitals throughout Japan (clinical registration with the University Hospital Medical Information Network: UMIN 000000736). The objectives and protocol of the ATTEND registry have been described in more detail previously and are summarized briefly here. Patients hospitalized for HF who met the modified Framingham criteria were eligible for the registry. Patients aged <20 years and those who were not considered suitable for inclusion in the registry by the attending physicians were excluded. The present study excluded patients with acute coronary syndromes. The data included in the present study were collected from April 1, 2007, and December 31, 2011. On admission, the patient’s history was obtained, and a physical examination was performed to assess New York Heart Association functional class, paroxysmal nocturnal dyspnea, orthopnea, rales, third heart sound, jugular venous distention, edema, coldness of the extremities, oxygen saturation by pulse oximetry, systolic and diastolic blood pressures, and heart rate. Echocardiography was also performed on admission to evaluate for a reduced left ventricular ejection fraction (≤40%). Blood chemistry tests were done on admission to measure blood urea nitrogen, serum creatinine, serum sodium, hemoglobin, C-reactive protein, total bilirubin, and brain natriuretic peptide (BNP). In-hospital mortality was defined as (1) death from any cause, (2) death from cardiac causes including sudden cardiac death and HF death, and (3) death from cerebral or vascular causes. Death was considered to be cardiac (defined as HF death, sudden death, or other cardiac death) unless a specific noncardiac cause was identified by each primary physician. The end point classification committee (2 experienced cardiologists who were not study investigators) reviewed the data and, if any problems were encountered, asked the primary physician to confirm the cause of death. Finally, the committee categorized each event for use in the present analysis.

The study was conducted on the basis of the Declaration of Helsinki and the Japanese ethical guidelines for clinical studies.

Of the 4,842 subjects registered, 5 patients with missing serum sodium levels on admission or incomplete data were excluded, leaving 4,837 patients for analysis. Analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, North Carolina). Data are presented as mean ± SD and as medians with interquartile ranges. Student’s t tests were used to compare groups with respect to normally distributed continuous variables, and Mann-Whitney U tests were used for other variables. Chi-square tests were used to compare nominally scaled variables. To evaluate the influence of hyponatremia with respect to in-hospital mortality, a logistic regression model was applied, and heterogeneity of the influence, or interaction, was evaluated in the model. Two-tailed p values <0.05 were considered to indicate statistically significant differences. All analyses were performed at an independent biostatistics and data center (STATZ Institute, Inc., Tokyo, Japan).

Results

In 4,837 hospitalized patients with HF, hyponatremia (<135 mEq/L) was observed in 11.6% (n = 561) ( Table 1 ). The distribution of serum sodium levels is shown in Figure 1 . Comparison of patient characteristics between patients with hyponatremia and those with normonatremia are listed in Table 1 . Compared to those with normonatremia, patients with hyponatremia were of similar age (73 years), included slightly fewer men, less commonly had hypertensive heart disease, and had higher proportions of history of hospitalization for HF and cardiac resynchronized therapy. On admission, heart rates and blood pressures were significantly lower in patients with hyponatremia than in those with normonatremia. Plasma BNP levels were significantly (p = 0.001) higher in patients with hyponatremia than in those with normonatremia. Impaired renal function and anemia were observed more commonly in hyponatremic hospitalized patients with HF.

Distribution of serum sodium levels on admission. The proportion of patients with hyponatremia (<135 mEq/L) was 11.6% (n = 561).

Before admission, angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker and calcium-channel blocker use was not different between patients with hyponatremia and those with normonatremia, but diuretic use was more common, especially thiazide diuretics and spironolactone, and β-blocker use was slightly higher in patients with hyponatremia ( Table 2 ). During hospitalization, inotropes use was significantly higher and vasodilator use significantly lower in patients with hyponatremia. Nonpharmacologic management, such as cardiac resynchronization therapy, implantable cardioverter-defibrillators, continuous hemodiafiltration, intra-aortic balloon pumps, percutaneous cardiopulmonary support, and left ventricular assist systems, were used more commonly in hyponatremic patients with HF, although the use of noninvasive positive pressure ventilation, intubation, and Swan-Ganz catheterization were not different between the 2 groups.

In patients with hyponatremia, a significantly longer length of hospital stay was observed ( Table 2 ). The all-cause death rate was significantly higher in patients with hyponatremia at 15.0% compared with 5.3% in those with normonatremia. All-cause death rates in patients with hyponatremia and those with normonatremia at 7 days after admission were 1.4% and 1.0%, respectively, and cardiac death rates were 1.4% and 0.8%, respectively. The 2 rates were not statistically different between patients with hyponatremia and those with normonatremia. Furthermore, cardiac death rates were also significantly higher in hyponatremic patients with HF, as shown in Figure 2 . In hyponatremic patients with HF, cardiac death accounted for 76.2% of all-cause death, higher than in normonatremic patients with HF (68.0%). In patients with hyponatremia who died from cardiac causes, systolic blood pressure at admission, known as a powerful independent risk factor for HF death, was significantly lower (107 ± 22 mm Hg) than in patients with noncardiac causes of death (137 ± 34 mm Hg) (p <0.001).

Relation between admission serum sodium level and in-hospital all-cause (left) and cardiac (right) deaths in all patients. Both causes of death were significantly (p <0.001) different between patients with hyponatremia and those with normonatremia. In Japan, the mean length of hospitalization is 30 days, quite different from that of Western countries. ^∗ p <0.001.

To evaluate the heterogeneity of the prognostic value of hyponatremia on admission with respect to all-cause and cardiac mortality, subgroup analysis was performed. As listed in Tables 3 and 4 , there were interactions for age, C-reactive protein, and BNP in all-cause and cardiac death. Also, an interaction existed for serum creatinine regarding cardiac death. Given these findings, age, renal function, C-reactive protein, and BNP should be considered from the viewpoint of the clinical significance of hyponatremia in cardiac death.

Table 3

Subgroup analysis of the prognostic value of hyponatremia in the prediction of all-cause death

Stratum	Hyponatremia (Serum Sodium <135 mEg/L)		Normonatremia (Serum Sodium ≥135 mEg/L)		Odds Ratio	95% Confidence Interval	p Value	p Value for Interaction
	Number of Patients	Number of Events (%)	Number of Patients	Number of Events (%)
Total	561	84 (15.0)	4,276	228 (5.3)	3.13	(2.39–4.09)	<0.001
Age (yrs)
≤74	249	30 (12.0)	2,007	46 (2.3)	5.84	(3.61–9.45)	<0.001	0.002
≥75	311	53 (17.0)	2,254	180 (8.0)	2.37	(1.70–3.30)	<0.001
Gender
Female	264	39 (14.8)	1,772	102 (5.8)	2.84	(1.91–4.21)	<0.001	NS (0.530)
Male	297	45 (15.2)	2,504	126 (5.0)	3.37	(2.34–4.85)	<0.001
Ischemic
Absent	396	62 (15.7)	2,929	140 (4.8)	3.70	(2.69–5.09)	<0.001	NS (0.086)
Present	164	22 (13.4)	1,339	88 (6.6)	2.20	(1.34–3.63)	0.002
History of HF
Absent	313	35 (11.2)	2,729	118 (4.3)	2.79	(1.87–4.14)	<0.001	NS (0.559)
Present	245	48 (19.6)	1,503	104 (6.9)	3.28	(2.26–4.76)	<0.001
Edema
Absent	165	24 (14.5)	1,380	71 (5.1)	3.14	(1.91–5.14)	<0.001	NS (0.964)
Present	392	59 (15.1)	2,843	150 (5.3)	3.18	(2.31–4.39)	<0.001
Left ventricular ejection fraction (%)
≤40	320	47 (14.7)	2,261	132 (5.8)	2.78	(1.95–3.96)	<0.001	NS (0.340)
>40	228	35 (15.4)	1,890	90 (4.8)	3.63	(2.39–5.51)	<0.001
Systolic blood pressure (mm Hg)
≤140	377	73 (19.4)	2,052	158 (7.7)	2.88	(2.13–3.90)	<0.001	NS (0.428)
>140	183	11 (6.0)	2,213	64 (2.9)	2.15	(1.11–4.15)	0.023
Serum creatinine (mg/dl)
≤1.5	371	51 (13.7)	3,259	137 (4.2)	3.63	(2.58–5.11)	<0.001	NS (0.060)
>1.5	190	33 (17.4)	1,017	91 (8.9)	2.14	(1.39–3.30)	<0.001
C-reactive protein (mg/dl)
≤1.0	252	32 (12.7)	2,761	81 (2.9)	4.81	(3.13–7.41)	<0.001	<0.001
>1.0	299	51 (17.1)	1,458	144 (9.9)	1.88	(1.33–2.65)	<0.001
BNP (pg/ml)
≤1,000	314	50 (15.9)	2,581	95 (3.7)	4.96	(3.44–7.14)	<0.001	<0.001
>1,000	210	31 (14.8)	1,382	119 (8.6)	1.84	(1.20–2.81)	0.005

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Comparison of Change in Coronary Atherosclerosis in Patients With Stable Versus Unstable Angina Pectoris Receiving Statin Therapy (from the Treatment With Statin on Atheroma Regression Evaluated by Intravascular Ultrasound With Virtual Histology [TRUTH] Study)
2. Left Ventricular Fibrosis in Atrial Fibrillation
3. Diagnostic Accuracy Retrospectively of Electrocardiographic Findings and Cancer History for Tamponade in Patients Determined to Have Pericardial Effusion by Transthoracic Echocardiogram
4. Erratum for Careri G, et al. “Clinical Correlates and Prognostic Value of Flow Mediated Dilation in Patients With Non-ST Segment Elevation Acute Coronary Syndromes.” Am J Cardiol 2013;111:51–57

Stay updated, free articles. Join our Telegram channel

Join