Introduction

Congestion is the hallmark of heart failure (HF) and therefore evaluation for congestion and its severity is the cornerstone in managing HF patients. Congestion not only leads to symptoms and signs of HF but is also linked to HF progression [1] and is associated with post-discharge morbidity and mortality [2]. It is known that many patients hospitalized with HF are not optimally decongested at time of discharge [3], leading to rehospitalization due to HF decompensation. In a recent analysis of the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial, Cooper et al. found that post-discharge outcomes of patients with HF were mainly driven by persistent congestion (higher right and left-sided filling pressures) [4]. Thus, the most reliable way to reduce HF admission and readmission is to monitor severity of congestion and optimally treat it.

The congestion cascade begins days to weeks prior to symptom onset with subclinical rise of left and right-sided filling pressures (hemodynamic congestion), which is followed by redistribution of fluids in the extravascular space and visceral organs (organ congestion), ultimately leading to symptoms and signs of volume overload (clinical congestion) [5]; therefore, it is prudent to recognize the onset of clinical congestion early enough to allow medication adjustment and/or hospitalization if necessary. Validity of physical signs in the diagnosis of HF decompensation has been acknowledged a long time ago [6], but the correlation of these signs with direct hemodynamic measurements has rarely been reported.

It is unknown whether subjective symptoms versus objective physical examination signs of volume overload are more important for the diagnosis of HF decompensation among patients with left ventricular systolic dysfunction. We have recently shown the value of a simple physical examination sign, the positive hepatojugular reflux (HJR), in its ability to identify the degree of congestion through its strong association with inferior vena cava diameter and central hemodynamic parameters of volume overload: the pulmonary capillary wedge pressure (PCWP) and right atrial pressure (RAP), in addition to its ability to predict post-discharge outcomes [7]. Herein, we aimed to provide a more detailed assessment of the value of history and physical examination in identying degree of congestion. In a recent study, the sum of PCWP + RAP was used as a measure of congestion as it reflects both right and left ventricular performance [8]. We utilized this criterion to grade the severity of congestion, and compare patient symptoms and signs of volume overload according to the sum of PCWP and RAP, in order to identify the best tools for evaluation of degree of congestion in HF.

Methods

This study is a retrospective analysis of a limited access dataset from the ESCAPE trial provided by the National Heart, Lung and Blood Institute (NHLBI) [9]. The ESCAPE trial enrolled 433 patients hospitalized with acute HF, and compared outcomes of groups of patients whose care was guided by pulmonary artery catheters (PAC) versus those managed by clinical assessment alone. The following were the study inclusion criteria: hospitalization for decompensated HF within the previous year, urgent visit to the emergency department, treatment during the previous month with 160 mg of furosemide or its equivalent, experiencing at least 3 months of HF symptoms despite being on angiotensin-converting enzyme inhibitors and diuretics, systolic blood pressure ≤125 mmHg, and left ventricular ejection fraction ≤30% in addition to at least one symptom and one sign of congestion. The study concluded that the use of PAC did not improve or worsen HF outcomes.

The objective of the current analysis was to identify the best history and physical examination components that predict the severity of HF exacerbation. The ESCAPE trial is ideal for such analysis as it comprised variables of the history and physical examination at multiple time points including hospital admission. Variables of HF decompensation related to patient history included dyspnea, orthopnea and fatigue. The study also contained detailed physical examination findings including the presence of rales, S3 gallop, jugular venous distension (JVD), lower extremity edema, HJR, hepatomegaly and ascites. These variables were also dichotomized according to the severity of history and physical examination sign (for example, dyspnea and fatigue were graded from 1–3 and orthopnea was graded from 1–5). We determined the severity of HF exacerbation in each patient by the sum of PCWP and RAP on admission, which assesses the degree of pulmonary and systemic congestion, respectively, and therefore only patients who were managed via PAC were included in the current analysis. Cases were divided into tertiles according to the sum of PCWP and RAP on admission. History and physical examination variables of congestion were examined across tertiles.

Statistical analysis

Primary analysis compared history and physical examination signs of congestion across tertiles of PCWP + RAP. Categorical variables were compared using the χ2-test and results were summarized as counts and percentages. Continuous variables were compared using the Kruskal-Wallis test because of non-normality of distribution (assessed using the Shapiro-Wilk test) and results were presented as median and interquartile range (IQR). The ability of the magnitude of the combined value of PCWP and RAP on admission to predict all-cause mortality and rehospitalization for HF, both at 6 months after randomization was assessed using receiver operating characteristics (ROC) curves. Optimum cut-off values of PCWP + RAP were those that provided the highest combined sensitivity and specificity for predicting the outcome. P-values for the trend across tertiles were examined via linear regression analysis for continuous variables and χ2-test linear-by-linear association for categorical variables. Other than ROC curve analysis which was performed via MedCalc software (MedCalc, Ostend, Belgium), the rest of the statistical analysis was performed using IBM SPSS 21.0 statistical software (IBM SPSS version 21.0. Armonk, NY). All statistical significance was assessed using a 2-sided P-values. A P-value < 0.05 was considered statistically significant.

Results

Baseline characteristics

A total of 187 cases that were randomized to receive PAC and had data available for both PCWP and RAP were included in our analysis. These cases were divided into tertiles according to the combined value of PCWP and RAP as follows: tertile 1: PCWP + RAP < 31 mm Hg, tertile 2: PCWP + RAP 31–42 mm Hg and tertile 3: PCWP + RAP > 42 mm Hg. There were significant differences across tertiles with regards to baseline B‑type natriuretic peptide (BNP) levels (P = 0.016), blood urea nitrogen (BUN, P = 0.022), sodium (P = 0.015), left ventricular ejection fraction (P = 0.005), left ventricular end-systolic dimension (P = 0.049), and the diameter of inferior vena cava during inspiration (P < 0.001) and expiration (P < 0.001). Comparison of baseline characteristics among patients enrolled in the ESCAPE trial according to tertiles of PCWP + RAP is listed in Table 1.

Table 1 Comparison of baseline clinical, laboratory and echocardiographic characteristics among patients enrolled in the ESCAPE trial according to tertiles of combined pulmonary capillary wedge pressure and right atrial pressure
Full size table

Best history and physical examination tools that predict the degree of congestion

With regards to patient history, there was no significant differences across tertiles of PCWP + RAP in the degree of dyspnea (grades 1–3), orthopnea (grades 1–5) or fatigue (grades 1–3). With regards to physical examination on hospital admission, there were significant differences across tertiles in the frequency of JVD (P < 0.001), or extent of JVD > 12 cmH2O (P < 0.001), the presence of lower extremity edema (P = 0.001) and lower extremity edema at least 2 + (P = 0.029), a positive HJR (P = 0.022) and the presence of ascitis (P = 0.009). There were no differences across tertiles of PCWP + RAP in direct signs of elevated left sided filling pressures specifically the presence of rales and S3 gallop. A comparison of symptoms and signs of volume overload across tertiles of PCWP + RAP is listed in Table 2.

Table 2 Comparison of history and physical examination findings among patients enrolled in the ESCAPE trial according to tertiles of combined pulmonary capillary wedge pressure and right atrial pressure
Full size table

Post-discharge outcomes according to admission PCWP + RAP

There was a significant difference across tertiles of PCWP + RAP with regards to 6‑month all-cause mortality (P = 0.029), rehospitalization for HF (P = 0.031), composite end-point of death, cardiac rehospitalization and cardiac transplantation (P = 0.004), composite end-point of death, any rehospitalization and cardiac transplantation (P = 0.022), and total duration in-hospital in the first 30 days (P = 0.001). Using linear regression and Chisquare linear-by-linear association, we found a significant increasing trend across tertiles of PCWP + RAP with regards to the aforementioned end-points. Comparison of post-discharge outcomes (and trends) across tertiles of PCWP + RAP is listed in Table 3.

Table 3 Short and intermediate term outcomes of patients enrolled in the ESCAPE trial according to tertiles of combined pulmonary capillary wedge pressure and right atrial pressure
Full size table

ROC curves showed that PCWP + RAP on admission had an AUC of 0.623 (95% CI 0.550–0.693, P = 0.0075) in predicting 6‑month all-cause mortality. An optimal cutoff criterion of PCWP + RAP ≥ 40 mm Hg had the highest combined sensitivity (63.4%) and specificity (63%) in predicting mortality (Fig. 1, panel a). Also, PCWP + RAP also had an AUC of 0.617 (95% CI 0.544–0.687, P = 0.0048) in predicting rehospitalization for HF, and an optimal cutoff criterion of PCWP + RAP > 38 mm Hg had the highest combined sensitivity (56.4%) and specificity (65.1%) (Fig. 1, panel b).

Fig. 1
figure 1

ROC curves showing the ability of the sum of pulmonary capillary wedge pressure (PCWP) and right atrial pressure (RAP) on hospital admission of patients with systolic HF in predicting death (panel a) and rehospitalization for heart failure (panel b), both at 6 months (AUC area under the curve)

Full size image

Discussion

We have shown in this analysis that certain physical examination signs of congestion such as JVD, a positive HJR, and the amount of lower extremity edema, determine the severity of congestion among patients with HF exacerbation. The JVD not only reflects increased right sided filling pressures but was also shown to be the main clinical sign to indirectly detect elevated leftsided pressures due to the high degree of concordance between right and leftsided chambers in over two thirds of HF patients [10]. JVD was also independently associated with higher short and long-term mortality [11]. The extent of lower extremity edema is one of the important signs used in monitoring and managing HF patients. Although the importance of lower extremity edema was minimized in the Framingham criteria for diagnosing HF due to its low specificity, it is more important in guiding therapy of already established HF cases. The HJR is an underutilized, simple physical examination sign with high intraobserver agreement of approximately 97% [12] that was found previously by our group to correlate well with other physical examination signs of congestion, BNP level, as well as the PCWP and RAP [7]. In our cohort, ascites was present at baseline with increasing frequency across tertiles of PCWP + RAP (22.4% of patients in first tertile of PCWP + RAP had ascites vs. 32.3% of patients in tertile 2 vs. 48.4% in tertile 3, P = 0.009). Those with ascites were most likely to have grade 1 (trace ascites, 20% of the cases included in our analysis), followed by grade 2 (moderate ascites, 13%) and less likely to have grade 3 (massive ascites, 1.6%); therefore, if present only modest amount of ascites is expected in HF exacerbations even with more severe congestion.

Not surprisingly, we did not find any value for conventional clinical signs of increased left sided filling pressure (e. g. rales and S3 gallop) in predicting the degree of congestion. It is known that patients with chronic HF often lack pulmonary rales despite elevated leftsided filling pressures due to chronic lymphatic hypertrophy, which prevents alveolar edema development despite high interstitial pressures [13]. Symptoms of volume overload, such as dyspnea, orthopnea and fatigue and their severity were also not associated with the degree of congestion. In fact, we have previously shown from an analysis of the ESCAPE trial that symptoms in HF correlate poorly with objective parameters of congestion [14]. Moreover, the ESCAPE trial enrolled patients with acute systolic HF with advanced symptoms and approximately 90% of patients included in our analysis had New York Heart Association class IV symptoms, thereby limiting the diagnostic value of symptoms of volume overload in predicting severity of congestion in this dataset.

We used the combined sum of PCWP and RAP as a measure of congestion as it represents volume loading of both the pulmonary and systemic compartments and is thus an index of biventricular performance [15]. A recent study illustrated the prognostic effect of PCWP + RAP obtained post-treatment on survival and readmission at 6 months [8], and our analysis similarly confirms the prognostic effect of this marker on hospital admission in predicting morbidity and mortality. The misperception that the physical examination provides limited information compared with modern diagnostic tools has led to a gradual decline in emphasis of the value of physical examination in HF. The significant association that we found between the presence and magnitude of JVD and lower extremity edema, a positive HJR, and PAC-derived intracardiac pressures, highlights the value of these physical examination tools in evaluating patients with HF, and to guide medication adjustment and decision for hospitalization.

Conclusion

We have shown the importance of certain physical examination signs in identifying the severity of HF exacerbation. Specifically, these are signs of more severe systemic congestion and include the extent of JVD, lower extremity edema and a positive HJR. Those signs correlated well with central hemodynamic markers of congestion measured via the PAC. Neither the physical examination signs of pulmonary congestion such as S3 gallop and the presence and extent of rales, nor HF symptoms, such as dyspnea, orthopnea or fatigue were valuable in determining the degree of congestion during HF exacerbation. The study further highlights the value of physical examination skills in acute HF.