Novel 4-Stage Classification to Estimate Right Atrial Pressure by Point-of-Care Ultrasound of Neck Vasculature

There has been recent interest in point-of-care ultrasound (POCUS) assessment of the internal jugular veins (IJVs) to estimate right arial pressure (RAP),^1,2 given that the IJV is superficial and more easily imaged than the inferior vena cava. Previous studies assessed changes in the IJV with respiration or the Valsalva maneuver,¹ as well as transition points in IJV size,² among other approaches. However, no method has been universally adopted.

Herein, we describe a novel 4-stage POCUS protocol that qualitatively compares the IJV and carotid artery (CA) areas, reasoning that the CA provides an anatomical reference for indexing IJV size and would not require estimation of the distance above the right atrium. We evaluated the protocol’s diagnostic utility for estimating RAP in patients with heart failure (HF), with a focus on detection of an elevated RAP given the clinical utility of that finding; pulmonary artery catheterization (PAC) was used as the reference standard.

METHODS

This was a cross-sectional single-center cohort study of patients with HF undergoing PAC for clinical indications. The Institutional Review Board at the University of Texas Southwestern approved the protocol. All participants provided written informed consent.

Between September 2024 and March 2025, a total of 136 patients (11 in a pilot phase) aged >18 years with a HF diagnosis (right- or left-sided) scheduled for PAC were enrolled (Figure 1). This was a convenience sample based on investigator availability. The indication for PAC was either left-sided HF or pulmonary hypertension. Patients with the latter were eligible if they had a documented history of right- or left-sided HF. Key exclusion criteria included intubation, noninvasive positive pressure ventilation, bilateral IJV indwelling catheters or thromboses, significant neck deformities, severe carotid stenosis or stents, ventricular assist devices, heart transplant, or pregnancy. After the pilot study, 8 participants were excluded for scheduling purposes (n = 4), lack of clear POCUS images (n = 2), a thyroid mass (n = 1), or intolerance of body positioning (n = 1). Exclusion occurred without knowledge of the PAC results. A total of 117 patients were included in the final study cohort.

FIGURE 1. A Novel 4-Stage POCUS Protocol for Estimating RAP Using Internal Jugular to CA Comparison.

The internal jugular vein (IJV) and carotid artery (CA) areas were assessed across several neck positions by using point-of-care ultrasound (POCUS). The highest of the stages from the right or left side was taken as the final stage. Mean right arial pressure (RAP) (error bars depict SD) increased across the 4 stages (P < 0.001). The P values depicted are for various comparisons of 2 of the 4 stages. The protocol had useful discrimination for elevated RAP. AUC = area under the curve.

POCUS was performed (D.d.O.-G., R.M.D., M.D.) using POCUS iQ3 or iQ+ probes (Butterly Network) on the same day and shortly preceding PAC. Initial images were obtained with the participant at an approximately 45° angle to the horizontal (Figure 1). The probe was placed at mid-neck, angled perpendicular to the skin, and the IJV and CA areas were compared qualitatively. If the maximum cross-sectional IJV area was larger than the CA area during most respiratory cycles, the IJV was then considered larger. Care was taken to apply only light pressure to avoid mechanical compression of the IJV. The probe was then moved to immediately above the clavicle, angled parallel to the floor. The IJV and CA areas were again compared, first on the right side and then left side of the neck.

Next, the participant was elevated to 60° to 90° to the horizontal. Images were acquired with the probe at mid-neck, angled perpendicular to the skin, and IJV and CA areas were compared. The Valsalva maneuver was used to engorge the IJV, when collapsed, to confirm its presence. If one side of the neck was inaccessible (eg, indwelling catheter), the unobstructed side was used. Short video clips (~10 seconds) were recorded, and, in some cases (<10) in the early study phases, they were reviewed with other study investigators when there was uncertainty about the relative IJV and CA sizes.

RAP was categorized into 4 stages (Figure 1): Stage 1: IJV <CA both mid-neck and clavicle at 45°; Stage 2: IJV <CA mid-neck and IJV ≥CA at the clavicle at 45°; Stage 3: IJV ≥CA mid-neck at 45° and IJV <CA at 60 ° to 90°; and Stage 4: IJV ≥CA mid-neck at both 45° and 60° to 90 °. The final POCUS stage was based on the higher stage of the right- and left-sided images.

RAP was recorded at steady state at end-expiration during PAC. A senior investigator (J.T.T), blinded to the POCUS results, reviewed the hemodynamic waveforms to determine the RAP based on the mean a wave. Echocardiographic data were included if performed within 3 months before or after enrollment (n = 73).

RAP was compared across the 4 POCUS stages by using the Jonckheere-Terpstra test; the Kruskal-Wallis test was used for follow-up comparisons of RAP between the 2 stages, not adjusting for multiple testing. The area under the receiver-operating characteristic (ROC) curve (C-statistic) for elevated RAP (≥10 mm Hg) with 95% CI in the overall cohort and selected subgroups was determined. Post hoc, the 4 stages were collapsed into 2 stages (grouping Stages 1 and 2 and Stages 3 and 4); the Kruskal-Wallis test was used to compare RAP between the 2 stages, and ROC analysis was performed for RAP ≥10 mm Hg. Comparisons of C-statistics between ROC analyses were performed by using the DeLong method.³ Standard diagnostic utility measures were calculated.

Statistical analyses were performed by using SAS version 9.4 (SAS Institute, Inc).

RESULTS

The cohort had a mean age of 61 ± 14 years and a mean BMI of 29.1 ± 9 kg/m²; 44% were female, 65% were White, and approximately two-thirds were outpatients. Indication for PAC was HF (n = 81) or pulmonary hypertension (n = 36). In the overall cohort, mean RAP was 9 ± 6 mm Hg, and 48 participants (41%) had RAP ≥10 mm Hg.

Most participants were Stage 1 (n = 61 [52%]), followed by Stage 2 (n = 29 [25%]), Stage 4 (n = 18 [15%]), and Stage 3 (n = 9 [8%]). Mean RAP increased across stages (P < 0.001) (Figure 1). RAP was lower in Stage 1 vs Stages 2 (P = 0.035), 3 (P = 0.015), and 4 (P < 0.001). RAP in Stage 2 was lower vs Stage 4 (P < 0.001). There were no significant differences in RAP between Stages 2 and 3 (P = 0.6) or Stages 3 and 4 (P = 0.06). Stages 3 and 4 combined had high positive predictive value (PPV) (85%) for RAP ≥10 mm Hg, and Stage 1 had high PPV (80%) for RAP < 10 mm Hg. Combining Stage 2 with Stage 1 reduced the PPV for RAP <10 mm Hg to 72%. The PPV of Stage 2 for RAP <10 mm Hg was 55%. The discrimination of the 4-stage protocol for RAP ≥10 mm Hg was very good (C-statistic = 0.785) in the overall cohort. Subgroup analyses revealed consistent discriminatory performance among patients referred for PAC for HF (C-statistic = 0.777) or pulmonary hypertension (C-statistic = 0.791), as well as those with HF with reduced ejection fraction (C-statistic = 0.782) or moderate or severe tricuspid regurgitation (C-statistic = 0.794).

In a post hoc analysis, RAP was associated with a binary POCUS classification (7 ± 5 mm Hg in Stages 1-2, vs 15 ± 5 mm Hg in Stages 3-4; P < 0.001). However, the C-statistic (0.711) of this 2-stage model was lower compared with that of the 4-stage model (P = 0.01).

DISCUSSION

We developed and validated a novel 4-stage POCUS protocol to estimate RAP. Our main findings were: 1) RAP increased stepwise across the 4 stages; 2) the protocol had useful discrimination (C-statistic = 0.785) for RAP ≥10 mm Hg; 3) Stages 3 and 4 had a high PPV (85%) for RAP ≥10 mm Hg, although additional study of stage 3 is needed given its low prevalence (n = 9); and 4) for RAP <10 mm Hg, Stage 1 had a PPV of 80%, but Stage 2 had limited diagnostic utility.

The level of discrimination of the proposed 4-stage POCUS protocol for elevated RAP is comparable to that by physical examination. For example, in the ESCAPE (Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness) trial conducted at leading HF centers with experienced clinicians, the C statistic for the latter was 0.74.⁴ Notably, in the current study, the POCUS protocol was performed by operators with variable experience, including a medical student and internal medicine resident, highlighting its potential for application by clinicians who are not specialists.

STUDY LIMITATIONS.

This was a single-center study with a small sample size, but in this proof-of-concept study, the protocol had clinical utility. The qualitative estimates of the IJV and CA areas may have led to misclassification. A quantitative approach may have higher diagnostic utility, but we wanted to develop a protocol that could be implemented at the bedside without off-line processing. Review of video clips when there was uncertainty regarding relative IJV and CA sizes in the early phase of the study may have introduced bias, although it typically involved classification differences of stages that were one level apart. One-quarter of participants were stage 2, a stage that had limited diagnostic utility. However, in the remaining cohort (75%), stage 1 had utility for low RAP and stages 3 and 4 for high RAP. Using the CA as an indexation method may be affected by age-related changes or vascular remodeling in the CA, potentially introducing error. Lastly, additional studies of this protocol are needed to test interobserver reproducibility, replicate its diagnostic utility in a larger cohort with diverse medical conditions (eg, obesity) and settings (eg, emergency department), evaluate its relative utility vs other POCUS techniques including inferior vena cava assessment, and test its utility in participants without heart failure.

CONCLUSIONS

A novel, 4-stage POCUS protocol based on comparing the qualitative IJV and CA areas showed a stepwise increase in RAP across stages and useful discriminatory capacity for an elevated RAP. These findings suggest this approach may have clinical utility for noninvasive RAP assessment in patients with HF.

What is the clinical question being addressed?

Can a 4-stage POCUS protocol comparing internal jugular vein vs carotid artery size assess right atrial pressure in patients with heart failure?

What is the main finding?

Invasively measured right atrial pressure increased stepwise across the POCUS stages, and the protocol had useful discriminatory capacity (C-statistic = 0.785).