Atelectasis

Last updated on 22 Jun 2026

Image credit

You've just intubated a critically ill patient in the emergency department, and you're feeling pretty heroic.

Sure, it took you a little while to navigate through the secretions and blood to find the larynx, and okay the cuff needed a bit of a wiggle to get through the cords, but you got there, and the smooth square wave of the end tidal CO2 signalled your success to the world.

You've even listened to both lungs to check you haven't slid the tube in too far.

So why are the sats still dropping?

What happens normally

You and your patient have both evolved for negative pressure ventilation.

Elastic lungs hang suspended from the ribcage, with the former trying to collapse down, and the latter trying to swing open.

When relaxed, they sit at equilibrium with the lungs held open at their functional residual capacity.

This set up is very efficient, because it only takes a relatively small force to tip the scales and pull the chest wall open.

This generates a very small negative intrathoracic pressure of 2 - 5 cmH2O, and the atmosphere does the rest of the work.

💡

Because why not use the natural ventilatory power of 1033 cmH2O that's just sitting there?

As well as the energy efficient mechanism of breathing, this set up has other benefits:

The negative intrathoracic pressure helps hold alveoli, small airways and pulmonary blood vessels open
It also helps improve venous return from the body and head to the heart

What happens when you intubate

Three things happen when we render our patients insensible to the traumas of the real world and unable to breathe for themselves:

They lie down (hopefully before they're unconscious)
They lose essentially all of their muscle tone
We switch to a positive pressure ventilatory set up once we're plugged into the ventilator

When you lie down, the physics of your thorax changes dramatically. In the same way a whale isn't designed to lie on a beach, we're not really designed to lie flat on our back.

To start with, the abdominal organs and fat press upwards on the diaphragm, making it harder for the lungs to distend caudally.

Secondly, all of the chest wall mass (and breasts if present) press directly down on the ribcage and reducing thoracic expansion further.

Thirdly, the delicate balance of 'I'll send the air to where you're sending the blood' that we call V:Q matching gets thrown out of whack, so the whole cardiorespiratory system becomes significantly less efficient.

But that's okay, we can still breathe.

Unless we can't, because someone's given us rocuronium.

The diaphragm now loses the heavily underappreciated tonic activity that was keeping the abdominal organs at bay, and moves north
The chest wall muscles also relax further, reducing the cross sectional area of the chest
The spinal muscles relax which increases thoracic spinal curvature

These three factors reduce FRC by up to 20% and that's before you consider the upper airway relaxation and inevitable obstruction.

This is a large part of the reason why anaesthetised patients require more inspired oxygen than sleeping ones, especially if they're obese.

So why is this a problem?

Well if you reduce the volume of the lungs down far enough, then you run into two main issues:

The surface area available for gas exchange is much smaller
If you start going below closing capacity then the small airways start to close off, in a process called atelectasis

Remind me what closing capacity is

This is the volume in the thorax at which the smallest airways start to close.

It's not the volume per se that's causing the closure to occur, it's just the volume that the thorax happens to contain when this process starts to occur.

A capacity is a sum of volumes, so closing capacity is:

Residual volume plus
Closing volume

These are semantic definitions with no real clinical relevance.

What matters is whether your patient's closing capacity is larger or smaller than their functional residual capacity.

If closing capacity is smaller than functional residual capacity, then the small airways will remain open when the lungs relax down to FRC and no further
If closing capacity is larger than the FRC, we have an issue
As the lungs relax down to FRC, the small airways will start to close and there will be loss of gas exchange, and large chunks of the lung will become very unhelpful shunt

When you lie flat on your back, closing capacity is lower than FRC until the age of 44, after which point you start to get small airway collapse in the supine position.

Standing up, CC is lower than FRC until the age of 70.

If you can spout this in a viva, then you're golden - you don't need more detail than this for an exam scenario.

So what's atelectasis?

It's Greek.

ateles = ‘incomplete’
ektosis = ‘extension’

And it has been used since the 1800s to describe areas of lung collapse that leads to areas of intrapulmonary shunt.

💡

Shunt = perfused but not ventilated, Dead space = ventilated but not perfused.

Surely my wonderful hypoxic pulmonary vasoconstriction response will save me?

Unfortunately volatile anaesthetic agents partially inhibit hypoxic pulmonary vasoconstriction, reducing the main defence against shunt.

Oops.

How does it happen?

In four main ways:

Obstruction
Compression
Absorption
Surface tension

Obstruction

If an airway gets obstructed and can't ventilate properly, then everything downstream of that obstruction will collapse as the gas inside is gradually absorbed
This can be a lobule, a lobe or an entire lung

Compression

Direct compression from outside the airway physically squishes the alveoli and small airways shut
This could be an intrathoracic mass (tumour, haematoma, abscess)
Or an extrathoracic problem (obesity, direct external pressure)

Absorption

All gases are not equal in their rates of absorption from the alveoli into the blood
Nitrogen is relatively slow to diffuse, while oxygen is far quicker
Therefore an alveolus full of pure oxygen will theoretically deflate much more quickly, and to a greater extent, than one with a decent amount of nitrogen in it

This is why some anaesthetists insist on preoxygenating with slightly less than 100% oxygen, to ward off the evil atelectatic spirits.

Surface tension

Any stressful children's birthday party will tell you that elastic spheres inherently want to collapse
Laplace's law tells us that smaller alveoli require a higher distending pressure to remain open than larger alveoli
The converse is also true, meaning smaller spheres are much keener to collapse
Alveoli have surfactant which is designed to help splint alveoli open and prevent this

So it follows that anything that makes alveoli smaller, or disrupts the layer of surfactant, is going to result in more atelectasis.

In anaesthesia, compression atelectasis from reduced FRC and absorption atelectasis from high FiO₂ are the biggest culprits.

Who is at risk?

Ah excellent, a risk factors question - break it down into patient risk factors, anaesthetic factors, and surgical factors.

Patient factors

Obesity
Neonate
Elderly
Pregnant
Smoker
Lung disease - ARDS or chronic lung disease

All of these patients have a larger closing capacity compared to their FRC and therefore are more prone to atelectasis.

Anaesthetic factors

All induction agents except ketamine
High FiO2 - effect increases above 50%
Ineffective ventilation

Surgical factors

Position - head down is worse than supine is worse than prone
Cardiothoracic surgery
Bypass
Laparoscopic surgery
Prolonged procedure

Why it's a problem

Well if you're trying to oxygenate a patient whose alveoli are squished shut then you're both going to have a bad time.

Instead of the diaphragm smoothly pulling the alveoli open evenly across the base of the lungs where they're most compliant, and allowing the atmosphere to gently trickle lots of lovely nitrogen (nature's CPAP) in, you're using a metal box to shove large volumes of oxygen in from the top under high pressure.

This means high-concentration oxygen distributes to the areas of the lungs that are most willing to receive it (i.e. not the closed-off bits) which not only neglects the collapsed areas, but also encourages them to close too.

Now seems like a good time to mention the four traumas of the alveolar apocalypse:

Barotrauma

Pressure physically damages the walls of the delicate small airways

Volutrauma

Excessive tidal volume causes overdistension of alveoli and small airways

This can occur even when airway pressures are not particularly high.

Atelectotrauma

Repeated squishing shut and reopening is bad for alveoli
Being open is fine
Being squished shut isn't great for gas exchange but the alveoli don't mind all that much
It's the repeated opening and closing that does harm

Biotrauma

Any sort of damage or physical stress will induce some degree of inflammatory response
This inflammation then causes further damage to the fragile walls of the alveoli and small airways

All of these cause direct harm to the respiratory engine, and that's on top of the fact that the patient is already not oxygenating as efficiently as they should be.

What to do about it

Can we not just make negative pressure ventilators?

Yup, and they look like this

No atelectasis or access to the patient, you can only choose one.

PEEP

Positive end expiratory pressure is like pinching off a balloon before it fully deflates. The residual gas inside the balloon exerts a pressure that holds it open, and alveoli are the same.

PEEP aims to push the functional residual capacity back up above closing capacity, and keep the airways open, reducing shunt and atelectotrauma.

Not enough PEEP - doesn't work
Too much PEEP - volutrauma and haemodynamic compromise due to reduced venous return

💡

5 cmH2O of PEEP prior to induction of anaesthesia can help reduce atelectasis.

FiO2

Logically speaking, leaving a little nitrogen available to splint open the alveoli if you're really worried about rapid airway collapse isn't a crazy idea.

These guys found that in 36 healthy young women, preoxygenation with 60% oxygen resulted in 97% less atelectasis, however they also desaturated twice as fast.

80% oxygen produced 87% less atelectasis, and desaturated 30% quicker.

So it seems it does genuinely have an effect, you just need to decide what scares you more - more atelectasis or less safe apnoea time.

Lung protective ventilation

This is a simple combination of reducing excessive pressure and volume inside the lungs of mechanically ventilated patients.

6 ml/kg predicted body weight tidal volume
Keep the plateau pressure below 30 cmH2O
Lowest FiO2 possible
You can also thrown in permissive hypercapnia if the patient is really unwell, but we'll talk about that another time

It seems to be a low cost, low risk and easy way to reduce the risk of ventilator associated lung injury.

What's driving pressure?

Driving pressure = Plateau pressure − PEEP

Increasing tidal volume will often improve oxygenation, but it does not prevent that cyclic opening and closing of alveoli that causes atelectotrauma.

Modern lung protective ventilation therefore focuses less on achieving large tidal volumes and more on minimising driving pressure.

My take home

I've written this specifically as my take home as I don't want anyone thinking I'm handing out practice-changing advice or suggesting you do it a particular way.

In my practice, I have decided that preoxygenation with the APL valve closed ever so slightly to generate around 5 cmH2O of CPAP (patients usually don't tolerate any more as they feel like they can't exhale) with an FiO2 of 90 - 95% is an adequate compromise that makes me feel like I'm warding off bad juju, both for atelectasis and safe apnoea time, without being too controversial.

If you completely disagree, please absolutely do comment or get in touch, we love a debate and are always happy to be taught something new.