The present study was designed to assess the role of jugular venous distension (JVD) as a predictor of short- and long-term mortality in a “real-life” setting. The independent association between the presence of admission JVD and the 30-day, 1- and 10-year mortality was assessed among 2,212 patients hospitalized with acute heart failure (HF) who were enrolled in the Heart Failure Survey in Israel (2003). Independent predictors of JVD finding in study patients included: the presence of significant hyponatremia (odds ratio [OR] 1.48; p = 0.03), reduced left ventricular ejection fraction ([LVEF] OR 1.24; p = 0.03), anemia (OR 1.3; p = 0.01), New York Heart Association III to IV (OR 1.34; p <0.01) and age >75 years (OR 1.32; p = 0.01). The presence of JVD versus its absence at the time of HF hospitalization was associated with increased 30-day mortality (7.2% vs 4.9%, respectively; p = 0.02), 1-year (33% vs 28%, respectively; p <0.001), and greater 10-year mortality (91.8% vs 87.2%, respectively; p <0.001). Consistently, interaction term analysis demonstrated that the presence of JVD at the time of the index HF hospitalization was independently associated with a significant increased risk for 10-year mortality, with a more pronounced effect among younger patients, patients with reduced LVEF, preserved renal function, and chronic HF. In conclusion, in patients admitted with HF, JVD is associated with specific risk factors and is independently associated with increased risk of both short- and long-term mortality. These findings can be used for improved risk assessment and management of this high-risk population.
Heart failure (HF) is a common clinical syndrome and a major cause of morbidity and mortality. Despite significant improvement in the management of patients with HF, rates of readmission and in-hospital mortality remain high. Therefore, a simple clinical marker may be useful for improved risk stratification in patients hospitalized with HF. Acute decompensated HF often presents as an acute elevation of cardiac filling pressures. Jugular venous distension (JVD) is an easy and simple test to perform during the physical examination, without the costs and difficulty of interpretation of more complex methods. Patients with JVD are more likely to be hospitalized and have pump failure–related death. In an attempt to predict the prognosis of patients admitted to the hospital with the diagnosis of acute decompensated HF, many models and scores have emerged. However, data regarding the prognostic implications of physical examination findings in this population are limited and conflicting. We hypothesized that elevated jugular venous pressure on physical examination, as manifested by JVD, can provide incremental short- and long-term prognostic information in patients admitted with acute HF.
Methods
Baseline characteristics and admission data of patients admitted with acute HF were obtained from the Heart Failure Survey in Israel (HFSIS 2003) survey database. The design and methods of the HFSIS have been described previously. Briefly, the survey was conducted during March 2013 to April 2003 in all 25 public hospitals in Israel. The study included 93 of the 98 internal medicine and 24 of the 25 cardiology departments in Israel at that time.
The survey enrolled 4,102 patients of which 2,302 with an acute HF event, defined as: de novo HF or, an acute exacerbation of chronic HF. After excluding patients without a valid JVD result (n = 90), the final analysis cohort for this study included 2,212 patients (96% of the acute HF cohort).
HF was diagnosed by the local survey physicians according to the following prespecified criteria: (1) clinical presentation (symptoms/physical examination), (2) chest x-ray, (3) echocardiography, (4) radionuclide study, and (5) cardiac catheterization. Results of echocardiography, radionuclide scintigraphy, and catheterization findings were obtained either during or within the 6 months before the index hospitalization, unless the patient had a recent cardiac insult.
JVD assessment
Recording JVD was a prespecified survey objective and was assessed in most subjects (96%). Physical examination was performed only by physicians according to recommendations, while patients laid supine with a 30-degree angle. The definition of positive JVD was right internal jugular vein distension or right external jugular vein when the internal jugular vein cannot be determined, >9 cm of water above the right atrium (or 4 cm from the sternal angle).
Anemia was defined as hemoglobin <11 g/dl and significant hyponatremia as serum sodium ≤130 mmol/dl on admission. Diabetes mellitus was defined by one of the following criteria: a history of diabetes mellitus obtained from medical records, admission blood glucose ≥200 mg/dl, or the use of antidiabetic agents (on admission or discharge). New York Heart Association (NYHA) functional class was determined according to functional status and symptoms before index hospitalization. Left ventricular ejection fraction (LVEF) was determined by echocardiography and defined as reduced when <50% or preserved ≥50%. Obesity was defined as a body mass index >30 measured on admission. De novo (new-onset) HF was defined as the absence of previous diagnosis of HF, or admissions due to HF in the past. Renal function was categorized using the Modification of Diet in Renal Disease formula for estimated glomerular filtration rate (eGFR). Systolic blood pressure (BP) was defined as reduced when admission values were below 140 mm Hg.
The end points of this study were 30-day, 1-, and 10-year all-cause mortality, obtained during follow-up, either from the database itself (hospital charts) or by matching patient identification numbers with the Israeli National Population Registry. Every effort was made to ensure accurate and reliable profiling data, which included standardizing HF and data validation definitions. Survey forms were completed by physicians, and data accuracy and consistency was checked by dedicated software.
For the univariate analysis, percentages were calculated for categorical variables and means with SD for continuous variables. The chi-square test in case of categorical variables, with continuity correction for 2 × 2 tables in case of dichotomous variables, and the Student t test, in case of continuous variables, was used for measuring the significance of differences between the patients with and without JVD across the baseline characteristics. The cumulative probability of 30-day and 10-year all-cause mortality in patients with and without JVD was graphically displayed according to the method of Kaplan and Meier, with comparison of cumulative events by the log-rank test.
Logistic regression modeling was used to identify important predictors for the detection of JVD. The model included the following prespecified covariates: acute de novo HF (vs chronic HF exacerbation), hyponatremia, anemia, reduced LVEF, ischemic etiology of HF (vs other causes), NYHA functional class (III to IV), age >75 years, reduced admission systolic BP, obesity, and previous diagnosis of diabetes mellitus, chronic obstructive pulmonary disease, and hypertension.
We evaluated prespecified events in a binary logistic regression model adjusted for age, sex, eGFR, NYHA functional class, and the presence (vs absence) of JVD funding to explore the independent association of JVD finding with in-hospital adverse events. The events included were stroke, pneumonia, renal function worsening (defined as ≥0.3 mg/dl increase in serum creatinine levels), new-onset atrial fibrillation or flutter, and life-threatening arrhythmia (defined as events of cardiopulmonary resuscitation or sustained ventricular tachycardia or ventricular fibrillation).
Independent predictors of the 30-day, 1-, and 10-year all-cause mortality outcome were evaluated using binary logistic regression modeling including the following covariates: age, NYHA functional class, sex, and JVD presence.
We applied multivariate Cox proportional hazards analysis to evaluate long-term mortality predictors. The following prespecified covariates, of established prognostic significance, were introduced using the best subset method: sex, reduced systolic BP, hyponatremia, JVD presence, third heart sound, anemia, age >75 years, renal dysfunction (GFR <60 ml/min/1.73 m 2 ), obesity, NYHA functional class III to IV, diabetes mellitus, decreased LV function, and acute HF. In a secondary analysis, we evaluated the association between JVD and 10-year mortality in patients’ subgroups categorized by age (younger and older than 75 years), HF with preserved EF (HFpEF) versus HF with reduced EF, de novo HF versus acute exacerbation of chronic HF, eGFR over and below 60 ml/min/1.73 m 2 , and hemoglobin over and below 11 g/dl, using interaction term subgroup analysis.
Ethics committee at each of the participating hospitals approved the study protocol.
All p values calculations were 2-tailed and were considered statistically significant if their value was ≤0.05. The statistical analyses were performed with IBM SPSS, version 20.
Results
The HFSIS survey included 2,212 hospitalized patients for congestive HF, of whom 1,395 (63%) had JVD on admission. The baseline demographic, clinical characteristics and hospitalization data of study patients are presented in Table 1 . The mean age of total survey population was 75 T10 years. Sex and cardiovascular risk factors were similar between the 2 groups. However, patients with JVD displayed several important differences in their clinical characteristics compared with those without JVD at the time of hospital admission, including older age, higher serum creatinine value, higher rates of anemia, increased frequency of a third heart sound, worse NYHA functional class, and a higher frequency of previous myocardial infarction. Consistently, these patients were more often treated with furosemide and digoxin ( Table 1 ).
| Variable | Jugular venous distension | p-value | |
|---|---|---|---|
| Absent (n= 817) | Present (n=1,395) | ||
| Age > 75 years | 50.8% | 56.2% | 0.015 |
| Male | 54.3% | 55.9% | 0.41 |
| Left ventricle ejection fraction % | 39±14 | 37±15 | 0.007 |
| Ischemic etiology of HF | 72.1% | 70.5% | 0.43 |
| Diabetes mellitus | 43.9% | 43.2% | 0.75 |
| Chronic obstructive pulmonary disease | 19% | 20.1% | 0.54 |
| Body mass index > 30 kg/m 2 | 22.9% | 24.2% | 0.46 |
| Serum sodium < 130 mmol/dl | 5.1% | 7.9% | 0.01 |
| Hemoglobin < 11 g/dl | 27.2% | 33.1% | 0.004 |
| Estimated glomerular filtration rate ≤ 50 ml/min/1.73m 2 | 46.3% | 56.9% | < 0.001 |
| De novo HF | 32.5% | 21.7% | < 0.001 |
| Third heart sound | 3.9% | 8.4% | < 0.001 |
| New York Heart Association functional class III-IV | 41.4% | 52.5% | < 0.001 |
| Beta-blockers | 47.3% | 48.7% | 0.5 |
| ACEI/ARB | 56.8% | 58.4% | 0.5 |
| Loop diuretics | 60.6% | 74.1% | < 0.001 |
| Aldosterone antagonists | 17% | 27% | < 0.01 |
| Digoxin | 12.5% | 18.5% | 0.02 |
Multivariate logistic regression analysis identified several independent predictors of JVD finding on physical examination at the index hospitalization, including: exacerbation of chronic HF versus a de novo event, hyponatremia, reduced LV function, anemia, higher NYHA functional class, and age >75 years. Conversely, ischemic etiology of HF, reduced systolic BP at admission, diabetes mellitus, chronic obstructive pulmonary disease, and hypertension were not identified as independent predictors of JVD ( Table 2 ). Notably, obesity was not associated with the absence of JVD finding on physical examination. Patients with JVD finding were 28% more likely to have acute worsening of renal function (odds ratio 1.28, 95% CI 1.08 to 1.53) and acute events of atrial fibrillation or flutter (odds ratio 1.29, 95% CI 1.11 to 1.50) after adjustment for age, baseline eGFR, sex, and NYHA functional class. Finding of JVD was not associated with increased risk of in-hospital stroke, pneumonia, and ventricular arrhythmia or resuscitation ( Figure 1 ). The Kaplan–Meier survival analysis for 30-day all-cause mortality demonstrated significantly worse outcomes in patients with JVD, including worse short-term, 30-day mortality (7.2% vs 4.9%; log-rank p value = 0.02), 1-year (28% vs 33%, respectively; log-rank p value <0.001), and a significantly higher cumulative probability of mortality at 10-year follow-up (91.8% vs 87.2%, respectively; log-rank p value <0.001 [ Figure 2 ]).
| OR | 95% CI | p-value | |
|---|---|---|---|
| Exacerbation of chronic vs. “de novo” HF | 1.5 | 1.1-1.9 | < 0.01 |
| Male | 0.92 | 0.76-1.12 | 0.4 |
| Serum Sodium < 130 mmol/dl | 1.48 | 1.03-2.12 | 0.03 |
| Left ventricle ejection fraction < 50% | 1.24 | 1.02-1.51 | 0.03 |
| Hemoglobin < 11 mg/dl | 1.31 | 1.07-1.59 | 0.01 |
| New York Heart Association functional class III-IV | 1.34 | 1.11-1.61 | < 0.01 |
| Ischemic etiology of HF | 0.97 | 0.78-1.19 | 0.75 |
| Age > 75 years | 1.32 | 1.09-1.6 | 0.01 |
| Body mass index > 30 kg/m 2 | 1.15 | 0.92-1.43 | 0.22 |
| Hypertension | 0.99 | 0.78-1.24 | 0.9 |
| Chronic obstructive pulmonary disease | 0.88 | 0.70-1.12 | 0.31 |
| Diabetes mellitus | 0.94 | 0.75-1.18 | 0.61 |
| Admission SBP < 140 mm Hg | 1.01 | 0.84-1.22 | 0.88 |
∗ Logistic regression model was further adjusted for additional HF etiology (ischemic vs nonischemic).
Consistent with the unadjusted results, multivariate analysis demonstrated that JVD finding was independently associated with a significant adjusted 40% (p = 0.04) increased 30-day all-cause mortality risk ( Table 3 ). Additional significant predictors of worse outcomes were: older age and worse NYHA functional class. HF etiology and sex were not significantly associated with short-term outcomes. Long-term outcomes demonstrated that patients with JVD finding had a 15% (hazards ratio [HR] 1.15, 95% CI 1.05 to 1.27; p <0.01) increased risk of all-cause 10-year mortality. Additional significant predictors of worse outcomes were: hyponatremia, exacerbation of chronic HF (compared with de novo HF), renal dysfunction, lower systolic BP, anemia, worse NYHA functional class, and age older than 75 years ( Table 4 ). Interestingly, the presence of a third heart sound was not associated with adverse outcomes.
