From sound to differential diagnosis
How do you get from a sound you hear to a differential diagnosis?
Short answerFirst describe the sound with the four questions. Then ask why the sound has changed: is it the source, the flow or the path? Leave the name of the disease until last.
Why should you describe the sound first?
In earlier lessons you saw how to listen to a sound and record it. Now we go from that description to a differential diagnosis.
"Crackles on the right" is a starting point, but it does not give a diagnosis. We use the same word in fibrosis, pneumonia, heart failure and bronchiectasis. No single feature makes the diagnosis. When features are combined, the likely mechanisms and the differential diagnosis narrow. So first describe the sound with the four questions used in every lesson.
In inspiration or in expiration? At the start of inspiration or at the end?
Musical or explosive? Fine or coarse? Long or short?
In one area or widespread? On one side or on both? At the bases or in the upper zones?
Does it recur in the same place with every breath? Does it change with coughing, a deep breath, position or treatment?
What does the underlying breath sound change?
An adventitious sound is always heard on top of a breath sound. This underlying sound changes what the adventitious sound means. For example, crackles in one area suggest consolidation when they sit on top of bronchial breathing. Bronchial breathing means that you hear a harsh sound there, like the one over the trachea. In a hyperinflated chest, where breath sounds are diminished everywhere, the same crackles do not mean consolidation.
Why has the sound changed?
Once you have described the sound, ask why it has changed. Three questions are enough, and each one rests on one of the three rules (Where does sound come from?).
Has the source changed?
A new sound arises: a wheeze, crackles, stridor, a squawk or a friction rub. This is the third rule.
Has the flow changed?
The sound becomes weaker or stronger. Shallow breathing, severe obstruction, a mucus plug or an endobronchial lesion weakens it. This is the first rule: no flow, no sound.
Has the path changed?
The sound reaches the chest wall differently. Consolidation carries sound. Effusion, pneumothorax, hyperinflation and a thick chest wall weaken it. This is the second rule.
Sometimes you answer "yes" to two questions at once. Pneumonia produces new crackles, so the source changes. Solid lung carries sound better, so bronchial breathing is also heard and the path changes too. In obstructive atelectasis, air does not reach the area, so the flow changes. The blocked bronchus also cuts off the path of the sound.
When you find a silent area, first check your technique, then percuss. Dullness suggests fluid or dense lung. Hyperresonance suggests air (When breath sounds disappear).
What do new basal crackles mean in a patient with COPD?
Now apply the two steps to one patient. A patient with known COPD presents with increasing breathlessness, orthopnoea, peripheral oedema and recent weight gain. Breath sounds are diminished and expiration is prolonged, but this picture is familiar from earlier examinations. What is new is crackles at both bases.
Describe the new sound with the four questions:
- When? The new crackles are heard from mid to late inspiration.
- What does it sound like? The new crackles are fine. The patient's known sparse, early coarse crackles are still there.
- Where? At both bases, symmetrical, in the dependent regions.
- What changes it? The crackles persist after coughing.
Then ask the three questions. Has the source changed? Yes, new crackles have appeared at the bases. Has the flow changed? Not clearly, because the diminished breath sounds and prolonged expiration were already present. Has the path changed? No, the underlying sound is the same as before. You would answer "yes" if there were bronchial breathing in the same place, or a new silent area at the bases, for example from an effusion.
So the change is in the source. Crackles at both bases, symmetrical and in the second half of inspiration, fit congestion or fibrosis. They do not change with coughing, which makes secretions a less likely explanation. Orthopnoea, oedema and weight gain point to pulmonary congestion. A different picture would take you in another direction:
| Finding | What it suggests | How to test it |
|---|---|---|
| Crackles at both bases that do not change with coughing. Orthopnoea, oedema, weight gain. | Pulmonary congestion | Listen again after diuresis. |
| Crackles in one area with bronchial breathing in the same place. Fever, purulent sputum. | Pneumonia | Arrange imaging. Listen again during antimicrobial treatment. |
| Coarse crackles that shift after coughing | Secretions collecting in the airways | Listen again after airway clearance. |
To work through this case step by step, see the New breathlessness in a patient with COPD case.
How do you test your working diagnosis?
The first examination does not confirm the diagnosis, but it decides the next step. Listen again after diuresis, airway clearance or antimicrobial treatment, and ask: is the new sound decreasing in the way your suspected cause predicts? If the finding persists despite appropriate treatment, reconsider your working diagnosis.
Getting stuck on the first label. For example, in a picture first labelled "heart failure", crackles may persist on one side despite diuresis. If there is also bronchial breathing in the same place, consider a local lung process.
The most valuable comparison is with the patient's own previous examination. So write your findings as answers to the four questions. The next doctor can then listen for the same sound (A good examination note).
What do you do when several changes occur together?
Textbooks describe the mechanisms one at a time, but in patients several are often present together:
- COPD and heart failure: new crackles in the dependent regions of a hyperinflated chest with diminished breath sounds.
- Infected bronchiectasis: bronchial breathing in one area, added to coarse sounds from secretions.
- Combined pulmonary fibrosis and emphysema: late inspiratory crackles at the bases, diminished breath sounds in the upper zones.
- Pneumonia with effusion: better transmitted sound over the consolidated lung, and marked loss of sound over the fluid.
In such a patient, interpret each area separately and ask the three questions again for each one. For example, in pneumonia with effusion, both the source and the path have changed over the consolidation. New crackles arise and sound is transmitted better. Over the fluid, it is mainly the path that has changed. A sound that shifts, appears or disappears after treatment tells you something different from a sound that always stays the same.
Which questions do the stethoscope and ultrasound answer?
Lung ultrasound shows the pleura and the lung just beneath it: fluid, consolidation, interstitial syndrome and pneumothorax. In critically ill patients, it can detect some structural abnormalities better than auscultation.
Auscultation, in contrast, tells you about airflow, airway calibre and sound transmission. It is quick and can be repeated anywhere. If the sounds are recorded, it also suits remote monitoring and follow-up over time. The two methods complement each other. Ultrasound does not show airway narrowing, and the stethoscope can miss a small effusion.
- First describe a sound with the four questions: When? What does it sound like? Where? What changes it?
- The underlying breath sound changes the meaning of an adventitious sound.
- Then ask why the sound has changed: Has the source changed? Has the flow changed? Has the path changed?
- In a known disease, the main clue is what is new compared with the patient's previous examination.
- The first examination sets up a working diagnosis, and repeat examination tests it. If the finding does not change as expected, reconsider your working diagnosis.
Going deeperresidents · specialists · numbers and sources
From bedside finding to measurement
Moving to measurement does not need a new clinical language. It needs familiar observations to be expressed in a repeatable way. A wheeze can be described by its duration, the share of the respiratory cycle it occupies, its dominant frequency and the number of tones heard at the same time. Crackles can be counted and their position within inspiration can be located. The duration and spectral content of the waveform help to separate fine from coarse crackles. The underlying breath sound can also be summarised by a few measures: flow-corrected intensity, spectral centroid and spectral slope. To these are added the energy balance between inspiration and expiration and the asymmetry between regions.
How much does a single finding tell you?
The Tromsø community study examined 3,684 adults aged 40 years or older. Bilateral early inspiratory crackles were strongly associated with COPD. However, in people with early crackles, the positive predictive value for COPD was only about 23% (Melbye et al., 2021). The details of the study are in the Describing crackles lesson.
In the same community study, wheezes or crackles were heard at one or more sites in about 28% of participants (Aviles-Solis et al., 2019). So a single adventitious sound tells you very little without context (The limits of normal).
When is a change meaningful?
Two recordings from the same patient differ even when the patient is stable. Airflow, posture, sensor position, contact and ambient noise never stay exactly the same. So you first need to know the variability expected from the measurement itself and from the person's usual fluctuation.
In 702 recordings from 52 patients with fibrotic interstitial lung disease (ILD), interobserver agreement on the presence of crackles was moderate (κ = 0.57). Agreement on change over 12 months was lower (κ = 0.42) (Sgalla et al., 2024). For most acoustic measures, thresholds for meaningful within-person change do not yet exist. A well-established personal reference may be more informative than a threshold derived from the population.
Comparison with ultrasound
In acute respiratory failure, lung ultrasound can show pleural fluid, consolidation, interstitial syndrome and pneumothorax with high diagnostic accuracy (BLUE protocol, Lichtenstein and Mezière, 2008). A study comparing auscultation, chest radiography and ultrasound in ARDS supports this (Lichtenstein et al., 2004). In critically ill patients, ultrasound may outperform auscultation for some structural abnormalities.
The most likely advantages of acoustic recording lie in longitudinal monitoring, remote monitoring and settings where imaging is not immediately available. For this, sensor position, recording conditions and quality control must become sufficiently standardised.
From measurement to biomarker
For an acoustic measure to become a biomarker, it must pass four steps: analytical validity, reproducibility, clinical validity and clinical utility. In respiratory acoustics, many candidate measures have analytical and associative evidence. Evidence from prospective external validation and on decision value is limited. The steps are described in detail in the The place of digital auscultation lesson.
See the reference list for sources.
Check yourself
3 short questions. The explanation opens when you choose an answer.