A man in his late seventies has had three days of fever and a wet cough. He is breathing 26 times a minute on room air, and I can hear crackles at the right base. Pneumonia is sitting at the top of my list, and the reflex most of us have followed a thousand times kicks in. Let's get a chest X-ray.
An hour later, the portable film comes back. The read says no focal consolidation.
So now I have a patient who sounds like pneumonia and a picture that says he does not. Do I believe my ears and my clinical judgment, or do I believe the film? Or do I pick up the ultrasound probe that has been sitting beside the bed the whole time?
That composite case is the question behind this Journal Club. The paper is Chavez and colleagues, a systematic review and meta-analysis of lung ultrasound for diagnosing pneumonia in adults, published in Respiratory Research in 2014. It gets quoted every time someone argues for the probe over the film, which is exactly why it deserves a hard look.
Why this paper matters
Pneumonia is common, and in frail patients it kills. Getting the diagnosis right, and getting it right quickly, decides who gets antibiotics, who gets admitted, and who goes home safely.
The chest X-ray has earned its place. It is fast, cheap, and familiar. But when you check it against CT, it misses a lot. A 2015 meta-analysis that compared the two tests head to head found a pooled chest X-ray sensitivity of 77 percent. A 2025 update that used CT as the reference standard for every patient put it at about 73 percent. That is roughly 1 in 4 pneumonias missed on the film.
To be fair to the X-ray, a small consolidation that only CT can see is not always one that changes management. But it does mean a negative film is weaker reassurance than most of us treat it as, and it usually arrives an hour after we ordered it.
When Chavez and colleagues published, guidelines did not recommend lung ultrasound as an alternative to X-ray or CT for pneumonia. This meta-analysis set out to fill that gap.
The paper, in PICO
Population. The authors screened 2,726 records across 6 databases and included 10 prospective studies with 1,172 adults. The mean age was 59, and 60 percent were men. Six studies enrolled patients in emergency departments or on medical wards, and 4 enrolled critically ill patients in the ICU. Every study came from a high-income setting: Italy, France, Greece, Hong Kong, and one multicenter European study.
Index test. Lung ultrasound, with scanning protocols that varied widely. Only half the studies examined at least 12 zones of the chest. Seven studies used highly skilled physicians, 2 used emergency physicians or general practitioners who had been trained to scan, and 1 did not report operator training. Only 3 studies defined expertise, and the bar was high: at least 100 prior scans in one, and more than 10 years of lung ultrasound experience in two others.
Reference standard. Not uniform. Five studies used clinical criteria plus imaging, 3 used chest CT for every patient, and 2 used CT only when the X-ray and the ultrasound disagreed. Sonographers were blinded to the imaging results but not to the clinical picture.
Outcomes. Pooled sensitivity and specificity, likelihood ratios, and the area under the receiver operating characteristic (ROC) curve.
The results
Start with absolute numbers. Picture 100 patients who truly have pneumonia: lung ultrasound found 94 and missed 6. Now picture 100 patients who do not: ultrasound correctly cleared 96 and falsely flagged 4.
The likelihood ratios are where this becomes useful at the bedside. Here is what a scan does to your pretest probability, using the 2014 numbers and the more conservative 2025 numbers discussed below.
The result held across subgroups. In the 3 studies that used CT for every patient, sensitivity was 94 percent and specificity was 99 percent. In the 2 studies where emergency physicians or general practitioners had been trained to scan, the figures were 93 and 92 percent, although those confidence intervals reach down into the mid-80s. Only 4 studies reported how long the scan took, and the longest was 13 minutes.
On its face, the story is compelling. But an area under the curve of 0.99 should make a clinician lean in, not relax.
Critical appraisal
What is good
The search was broad, two reviewers independently extracted the data, and the QUADAS-2 assessment rated nearly every domain as low risk of bias. The ultrasound was interpreted blind to the X-ray and CT. The authors declared no competing interests, and the accuracy held up against their strongest reference standard.
Problem 1: This paper never tested the X-ray
The claim that ultrasound beats the film does not come from Chavez. It comes from setting these numbers beside X-ray numbers from other studies. The head-to-head evidence came later, and it is covered below.
Problem 2: 32 patients became 384 data points
Two studies counted lung regions or hemithoraces rather than people. One enrolled 32 patients and scored 12 regions in each, producing 384 observations with zero false positives. That single study is under 3 percent of the patients but about a quarter of all the observations in the pooled table. Regions within the same chest are not independent, so this inflates precision and pulls the pooled estimates toward that study's near-perfect result.
Problem 3: the rule-in wobbles, the rule-out holds
The positive likelihood ratio had a confidence interval from 7.7 to 37, with significant inconsistency between studies. The negative likelihood ratio, 0.05 to 0.10, was consistent. In this data set, the rule-out is the more trustworthy half of the test.
Problem 4: Who was in the room
Some included studies excluded patients with heart failure, interstitial lung disease, severe immunosuppression, suspected aspiration, and pregnancy. Those are the patients who make undifferentiated dyspnea in the ED genuinely hard. On the operator side, 7 of 10 studies used experts, and the authors themselves say that the 2 studies involving trained physicians are not enough to recommend lung ultrasound for non-experts. The only training program described was 3 hours of lectures and 3 hours of hands-on scanning.
What happens when you fix some of this
A 2025 meta-analysis of 8 studies used CT as the reference standard for every patient and compared both tests head-to-head. The numbers came down, but the direction held.
So here is the appraisal. The 2014 numbers look like a ceiling rather than an everyday expectation, but the direction has held up in better-designed work. Lung ultrasound, in trained hands, is a better test for pneumonia than the chest X-ray. The asterisk still reads "in trained hands."
A visual walkthrough of the findings
These are the findings to know, in the order you usually meet them. For moving examples, the LITFL lung ultrasound library by Dr James Rippey is an excellent free resource.
Lung sliding and A-lines: normal aerated lung. The pleural line shimmers as the lung moves beneath it, and evenly spaced horizontal lines repeat below it. A-lines are reverberation artifacts of the pleural line, and they tell you there is air under the probe.
B-lines: a sign, not a diagnosis. B-lines are bright vertical artifacts that start at the pleural line, run to the bottom of the screen, and move with lung sliding. A focal cluster can appear with pneumonia. Diffuse bilateral B-lines should point you toward pulmonary edema until proven otherwise. B-lines on their own are not pneumonia.
Consolidation and hepatization. When alveoli fill with fluid and inflammatory cells, the lung stops behaving like air and starts looking like solid tissue, often with a texture similar to liver. That is why you will hear the term hepatization.
The shred sign. Where consolidated lung meets aerated lung, the deep border is ragged and irregular, like torn paper. That fractal edge helps separate a consolidation from a pleural effusion, which has a smooth deep border.
The dynamic air bronchogram. Bright branching dots and lines inside a consolidation are air in the bronchi. When they move with each breath, that is a dynamic air bronchogram. In a study of 68 ventilated ICU patients whose consolidations contained air bronchograms, the dynamic version had a specificity of 94 percent for pneumonia over resorptive atelectasis. But its sensitivity was only 61 percent, so seeing it helps you rule in, and not seeing it does not rule out.
Pleural effusion. An associated effusion appears as a dark, fluid-filled space above the diaphragm and adds to the picture, although it is not specific to pneumonia.
How should this modify your practice?
Here is my position. I bring the probe to the bedside for any adult where pneumonia is on the differential, and I scan before the film comes back, not after. For the man in bed 12, a right base consolidation with dynamic air bronchograms, in that clinical picture, is pneumonia, and I treat it whatever the portable film says.
When the scan is clean across a complete protocol and my pretest probability is low to moderate, I lean away from pneumonia and look for the alternative diagnosis. Using the more conservative 2025 numbers, a 30 percent pretest drops to about 5 percent. That is a strong nudge, not a get-out-of-jail card.
Here is what I would not change. I still get a film when I am looking for something the probe does poorly, such as a mass, when the admitting team needs one, or when my scan was limited by body habitus, subcutaneous emphysema, or a patient who cannot sit forward for the posterior zones. And I do not call pneumonia on B-lines alone.
Here is who sits outside this evidence. Children, who have their own literature. And the patients some of these studies excluded: heart failure, interstitial lung disease, the immunocompromised, and pregnancy. Ultrasound is attractive in pregnancy because there is no radiation, but this paper cannot speak for those patients.
For trainees, this is a permission slip with a condition attached. Nobody yet knows how much training is enough. So do the course, save your clips, get them reviewed, and check your scans against the CT or the film whenever you can. That is how you earn the right to trust the probe.
And when you explain the scan to a patient, tell them it is not a shortcut. In trained hands, it is more sensitive than the X-ray, and they never have to leave the bed to get it.
The takeaway
Lung ultrasound, in trained hands, is a better pneumonia test than the chest X-ray, and a clean scan is the most trustworthy thing it gives you. The 2014 numbers are the ceiling, and building the skill is the whole ballgame.
And upstream of all of it: the best pneumonia is the one that never arrives. Make sure your patients have a primary care physician who keeps their vaccines up to date.
Watch or listen to the full episode, Overheard Journal Club Episode 8, on YouTube and wherever you get your podcasts.
About the author
Adrian Cois, MD, is an Emergency Physician and Assistant Professor of Emergency Medicine, and the host of Overheard in the Emergency Room. His philosophy is "be prepared, not scared": evidence-based medicine translated for clinicians, trainees, and patients.
Disclosure: Dr Cois has no financial relationships with any ultrasound manufacturer or with any resource linked in this post. There are no affiliate links. This post is for educational purposes only and does not provide medical advice or establish a physician and patient relationship.
Related reading
• Overheard Journal Club Episode 7: bedside ultrasound for renal colic and intranasal ketorolac
• Overheard Journal Club Episode 3: video vs direct laryngoscopy
• LITFL lung ultrasound library, Dr James Rippey: https://litfl.com/lung-ultrasound-pneumonia/
References
1. Chavez MA, Shams N, Ellington LE, et al. Lung ultrasound for the diagnosis of pneumonia in adults: a systematic review and meta-analysis. Respir Res. 2014;15(1):50. doi:10.1186/1465-9921-15-50
2. Ye X, Xiao H, Chen B, Zhang S. Accuracy of lung ultrasonography versus chest radiography for the diagnosis of adult community-acquired pneumonia: review of the literature and meta-analysis. PLoS One. 2015;10(6):e0130066. doi:10.1371/journal.pone.0130066
3. Vera-Ponce VJ, Ballena-Caicedo J, Bustamante-Rodríguez JC, et al. Pulmonary ultrasound versus chest radiography in the diagnosis of community-acquired pneumonia in adults: an updated systematic review and meta-analysis. Respir Med Res. 2025;88:101200. doi:10.1016/j.resmer.2025.101200
4. Lichtenstein D, Mezière G, Seitz J. The dynamic air bronchogram. A lung ultrasound sign of alveolar consolidation ruling out atelectasis. Chest. 2009;135(6):1421-1425. doi:10.1378/chest.08-2281
5. Rippey J. Lung ultrasound: pneumonia. Life in the Fast Lane (LITFL). https://litfl.com/lung-ultrasound-pneumonia/ *
