(Updated 2026)

From the stethoscope to the dialysis chair: what lung ultrasound can tell us, and where its limits lie.

When we suspect pulmonary congestion, we usually start with familiar clues: symptoms, crackles, peripheral edema, and chest radiography. But how reassured should we be when the lungs sound clear? Lung ultrasound may detect congestion that these traditional approaches miss. The more difficult question is whether acting on those findings actually improves outcomes. Finding more congestion and knowing how best to treat it are not quite the same thing.

It is easy to see why lung ultrasound (LUS) is attracting attention. It is quick, repeatable, does not involve radiation, and is remarkably sensitive to changes in lung aeration. For nephrologists, who regularly balance the risks of removing too little fluid against removing too much, this is particularly appealing. Still, we need to understand both what LUS shows and what it cannot tell us.

A clear chest does not mean clear lungs

The stethoscope is still useful, but it is not a very reliable way to screen for pulmonary congestion. In a study of 926 ICU patients, Cox and colleagues found poor agreement between auscultation and lung ultrasound. Among patients whose ultrasound showed pulmonary edema, roughly half had no abnormal findings on auscultation. The broader evidence points in the same direction. A meta-analysis by Arts and colleagues found that lung auscultation had an overall sensitivity of only 37% across acute pulmonary conditions. Crackles alone also performed only modestly for heart failure. By comparison, meta-analyses of LUS B-lines in acute cardiogenic pulmonary edema have generally reported sensitivities around 90% or higher. Those estimates do vary with the patient population and the scanning protocol, as Al Deeb and colleagues noted. So, a patient can have important pulmonary congestion without any crackles. This is especially relevant in hemodialysis, where considerable sonographic congestion can be present despite few symptoms or little peripheral edema, as illustrated by studies from Zoccali and colleagues and Torino and colleagues.

How does lung ultrasound compare with chest radiography and CT?

LUS often detects findings that chest radiography misses, especially when the radiograph is obtained with the patient lying supine. In adults with suspected acute decompensated heart failure, a meta-analysis by Maw and colleagues reported a sensitivity of 88% for LUS, compared with 73% for chest radiography, with broadly similar specificity. Another synthesis in critically ill patients also favored ultrasound over chest radiography when CT was used as the reference standard (Winkler and colleagues).

Other thoracic conditions show both the strengths and the limitations of this approach. For pleural effusions, ultrasound can pick up small fluid collections and show septations that may influence drainage decisions (Hansell and colleagues). With pneumonia, it can reveal subpleural consolidation that is missed on a radiograph. However, abnormalities deep in the lung that do not extend to the pleura remain beyond its view (Brogi and colleagues).

Pneumothorax is another good example. In trauma patients, pooled studies show that LUS is substantially more sensitive than supine chest radiography (Cochrane review). However, a negative limited scan does not exclude every small or anatomically hidden pneumothorax. CT remains important when clinical suspicion persists (Smit and colleagues).

For these reasons, it is a stretch to present LUS as a replacement for every chest radiograph or CT scan. However, if the main goal is to monitor whether cardiogenic pulmonary edema is improving, whether in the ICU or dialysis unit, there is really no reason to expose the patient to unnecessary radiation.

Detecting congestion is useful. But does treating it improve outcomes?

For nephrologists, this is probably the most important part of the discussion. B-lines are an ultrasound pattern associated with increased extravascular lung water. Observational studies show that greater sonographic pulmonary congestion in hemodialysis patients is linked to worse outcomes even in asymptomatic patients. In a cohort studied by Zoccali and colleagues, very severe congestion was associated with a markedly higher adjusted mortality risk. But identifying a risk marker does not prove that removing more fluid will improve survival.

The LUST trial helps put that distinction into perspective. It randomized 367 hemodialysis patients at high cardiovascular risk to LUS guided management or usual care. Lung congestion improved more often with the ultrasound strategy (78% versus 56%), without a clear increase in adverse events. However, the primary composite outcome of death, nonfatal myocardial infarction, or decompensated heart failure was not significantly improved (hazard ratio 0.88; 95% confidence interval 0.63 to 1.24). Post hoc analyses suggested possible benefits for recurrent cardiovascular events, but those findings should be considered hypothesis generating.

Smaller randomized studies have shown encouraging physiological effects. In hypertensive hemodialysis patients, Loutradis and colleagues found that LUS guided dry weight adjustment improved ambulatory systolic blood pressure. An echocardiographic substudy also reported improvements in selected measures of filling pressure and chamber size, with 12 month follow up providing further observations. These findings are meaningful, but they should not be mistaken for established reductions in mortality or major cardiovascular events. Systematic review evidence has likewise not established a convincing long term outcome benefit.

There is also a physiological reason not to turn B-line counts into an automatic ultrafiltration target. Pulmonary congestion does not necessarily mean that total body volume is excessive. Lung water reflects the combined effects of cardiac filling pressures, pulmonary vascular forces, permeability, and fluid balance. More B-lines may call for closer evaluation, but they do not, by themselves, establish intravascular volume, fluid responsiveness, or tolerance of additional fluid removal. A patient can have pulmonary congestion and still be unable to tolerate more ultrafiltration.

A more useful approach is to interpret LUS alongside the rest of the clinical picture: symptoms, blood pressure trends, interdialytic weight changes, cardiac findings when appropriate, and the response to previous fluid removal. The goal is not simply to generate a B-line count. It is to make a better clinical decision.

Do we really need to scan 28 zones?

Comprehensive 28 zone protocols have been valuable in research. But should we use the same approach for everyone in a busy dialysis unit? Assessing one critically ill patient is very different from evaluating dozens of patients sitting in dialysis chairs during a limited treatment window. In fact, investigators in the LUST trial acknowledged that many nephrologists found the technique complex and time consuming, which contributed to slow recruitment. While this does not necessarily imply poor scan quality, it highlights an important point: an examination can be technically excellent but still be impractical for routine use.

Fortunately, abbreviated protocols may work well for serial assessments of congestion. In 303 hemodialysis patients, Torino and colleagues found a very strong correlation between 8 site and 28 site lung ultrasound scores (ρ = 0.93), with similar prognostic performance. Median acquisition time fell from 3.05 to 1.35 minutes. In a separate emergency department study of 98 hemodialysis patients, Reisinger and colleagues found that 4, 6, and 8 zone protocols each correlated well with the 28 zone reference for grading congestion.

Of course, fewer scanning sites do not mean that every abbreviated protocol performs equally well in every situation. The locations of the scanning windows matter, especially when abnormalities are focal or lie in dependent lung regions. More comprehensive examinations may identify mild or regional abnormalities that a shorter scan misses, a tradeoff discussed by Leidi and colleagues and Mayr and colleagues. Even so, an 8 zone approach is a reasonable practical starting point for many serial congestion assessments, provided the examination is standardized and performed by trained clinicians.

The bigger question: what can we realistically implement?

There is a big difference between using LUS for a challenging inpatient case and making it part of routine care for everyone in an outpatient dialysis unit. In the hospital, a trained clinician (nephrologist/critical care doc) may have the time and expertise to complement LUS with focused echocardiography, venous Doppler, and other hemodynamic assessments. But doing this for dozens of patients in a busy dialysis unit is a different story. A shorter, reproducible protocol may therefore be more useful than a comprehensive examination that is difficult to implement consistently. Training dialysis nurses could make this more feasible, but that raises another practical question: will they be compensated for the additional work? Probably not! And regardless of who performs the examination, LUS still tells us only part of the story. It cannot independently assess forward flow, cardiac filling pressures, systemic venous congestion, or tolerance of ultrafiltration. These questions require additional clinical assessment, sometimes complemented by other ultrasound findings.