Lung ultrasound: monitoring congestion in patients with heart failure

Patients who have been hospitalized for acute heart failure (AHF) represent a particularly high-risk subset of heart failure (HF) patients, with approximately 50% being readmitted for HF or dying within the subsequent 6 months.¹ Irrespective of left ventricular ejection fraction, the management of congestion in patients following a hospitalization for AHF is essential in order to prevent early readmission. Current methods to assess congestion in the outpatient setting include monitoring patients’ symptoms, weight, and findings on physical examination. However, signs and symptoms are insensitive and qualitative, and the added problem of obesity and other co-morbidities in a rapidly aging population make the monitoring of congestion even more challenging.

Several observational studies and one randomized clinical trial have demonstrated that lung ultrasound (LUS) can be used to identify pulmonary congestion in patients with undifferentiated dyspnoea with higher accuracy than the clinical examination and chest X-ray.^2,3 In addition, LUS findings suggestive of pulmonary congestion (so called ‘B-lines’) identify individuals at high risk for decompensation or death. This has been shown in both ambulatory HF cohorts and in hospitalized patients with persistent B-lines at discharge.^4–6 A key unanswered question has been whether knowledge of LUS-detected lung congestion might lead to improved congestion management and, as a result, better outcomes.

The LUS-HF trial is, to our knowledge, the first randomized clinical trial using LUS to guide decongestive therapy in patients following a hospitalization for AHF.⁷ In this proof-of-concept study, 123 participants were randomized to either LUS-guided management or usual care prior to hospital discharge and followed for 6 months. Patients in both randomized groups underwent LUS with a pocket-size device. While patients were blinded to group assignment, treating clinicians were not but LUS results were withheld in the usual care group. Although there was no prescribed treatment protocol in the LUS group, treating clinicians were encouraged to increase decongestive therapy if the sum of B-lines in eight chest zones exceeded three B-lines. The primary endpoint was a composite of urgent visits for HF, HF hospitalizations, or all-cause death. Safety endpoints included hospitalizations for hypotension, hypo−/hyperkalaemia and worsening renal function. Several secondary endpoints were assessed, including N-terminal pro-B-type natriuretic peptide, B-line number, quality of life, and 6-min walk test. The LUS-guided HF management resulted in a reduction in the primary endpoint [hazard ratio (HR) 0.52, 95% confidence interval (CI) 0.27–1.00, P= 0.049] due to a reduction in urgent visits for worsening HF (LUS: 5% vs. control: 21%; HR 0.21, 95% CI 0.06–0.74, P= 0.015). In addition, there was an improvement in the change of the 6-min walk test distance in the LUS group [LUS: 60 m (29–125) vs. control: 37 m (5–70), P= 0.023], but no reduction in natriuretic peptides, B-line number or improvement in quality of life as compared to usual care. The risk of adverse events was similar in the two groups, which supports the safety of this approach.

How should we interpret these findings? Phase II trials investigating new medications typically use surrogate ‘efficacy’ endpoints, such as laboratory biomarkers, haemodynamic measures and assessment of symptoms/quality of life and functional capacity, with sample sizes commonly in the 200 patient range. Patient outcomes (e.g. HF hospitalizations and mortality) used in larger Phase III trials are often collected as well, although with little expectation of demonstrating any definitive effect because of the small sample size and short duration of follow-up. So, although the present findings are encouraging, with a signal of a potential reduction in urgent visits for HF, the results must be considered preliminary and not robust, as the authors appropriately highlight in their limitations section. The inclusion of patients with HF with preserved ejection fraction (HFpEF) in this trial is intriguing. In the absence of other evidence-based treatments to improve outcomes in this HF phenotype, managing congestion effectively may represent an important therapeutic strategy for HFpEF. While the patient cohort was similar to other HF trials with elderly patients with a broad spectrum of co-morbidities, including chronic obstructive pulmonary disease, patients were relatively slim (body mass index 26–27 kg/m²).⁸ In a cohort of HF patients with a higher prevalence of obesity the detection of B-lines with first generation pocket ultrasound devices may prove more challenging. As the state-of-the-art of pocket devices is changing rapidly, ultrasound systems with better image resolution could require the application of different B-line number thresholds in such patients. In addition, the real-time quantification of B-lines by investigators not blinded to group assignment or patients’ overall clinical appearance may have introduced bias. Use of standardized off-line image analysis in a core laboratory could reduce this potential bias in future trials.⁹ Nevertheless, the results of this proof-of-concept study warrant further investigation in larger, well-designed trials. Several such trials are currently ongoing or planned (e.g. NCT03613779, NCT03136198).

It is important to consider how LUS performs compared to alternative strategies used to achieve the same therapeutic goals. A number of biomarkers and devices for the detection and quantification of congestion in patients with HF have been investigated previously. These include natriuretic peptides (GUIDE-IT), pulmonary artery pressure monitors (CHAMPION), and devices measuring intrathoracic impedance.^10–12 Of these, natriuretic peptides are the easiest to measure and while useful for diagnostic and prognostic purposes, their utility in guiding HF therapy has not been convincingly demonstrated.¹⁰ On the other hand, treatment augmented in response to the results of pulmonary artery pressure monitoring has been shown to reduce HF hospitalizations in ambulatory patients with HF.¹¹ However, these devices are invasive and costly, meaning that they are not available in all regions of the world. Conversely, LUS maybe a universally feasible and cost-effective tool for monitoring pulmonary congestion allowing clinicians to more accurately adjust decongestive therapy in both patients with HF with reduced and preserved ejection fraction, but that possibility remains to be proven. Larger, well designed trials using LUS to guide management of patients with HF in a variety of settings (prehospital, emergency department, in-hospital, outpatient) are needed to better understand B-line cut-off values, identify the patients that are most likely to benefit and describe the implementation strategies needed in clinical practice, should this technique prove to be useful.

While the findings of this proof-of-concept study are encouraging, it is important to keep in mind that LUS remains one piece of the puzzle in the assessment of congestion in patients with HF. The results of this – and any other test – should not be considered in isolation but rather in the context of the patient’s history, physical examination and other investigations. For example, patients with predominant right-sided HF may not demonstrate any pulmonary congestion and others (e.g. those with interstitial lung disease) may demonstrate B-lines for reasons other than pulmonary congestion. The results of the LUS-HF trial should motivate the funding of larger randomized clinical trials to provide robust and definitive evidence as to whether LUS used to aid the management of patients with HF can improve clinical outcomes.