Unfit for surgery is not a plan
What happens to the air leaks we cannot operate on

A new pleural standard puts a clock on the call to a thoracic surgeon about a prolonged air leak. For the patients whose leaks last longest, the answer to that call is usually no — and nothing is started when it is.

When the British Thoracic Society published its pleural quality standard in May, I wrote here about the three statements that put a clock on the call to a thoracic surgeon. One of them concerns prolonged air leak: a pneumothorax still leaking after three to five days on a drain should prompt a discussion with thoracic surgery within twenty-four hours.

I welcomed it then and I still do. But there is a sentence in the standard I have kept returning to since. Surgery, it says, is the best option for sealing a leak and reducing the chance of a recurrence, while the alternatives — endobronchial valves among them — are less effective and are reserved for patients unfit to operate.

Every clause of that is true. But reserved is an odd word to use about a treatment, because it suggests something is being held ready. On a ward it usually means nothing further gets organised.

The answer to the twenty-four-hour question is usually no

The standard is written as though the difficulty is getting a surgeon into the conversation quickly enough. For the patients whose leaks go on longest, that is not the difficulty. The difficulty is that the surgeon's answer is no.

A leak persists because the lung will not seal, and lung that will not seal is diseased lung. Severe emphysema. Fibrosis. A fortnight of bacterial pneumonia that has hollowed out a lobe. The patients whose air leaks run into a third and fourth week are, more often than not, the patients in whom a general anaesthetic and single-lung ventilation carry a real chance of not being survived. The same pathology produces both facts, which means the group that waits longest and the group we cannot operate on are largely the same group.

The referral is made on time. I see the patient and explain that an operation would carry a real chance of killing him. The standard has been met in full. He is still attached to the drain, and nothing has been agreed about what happens next.

What follows is not a decision anyone makes

It is the absence of one. The team continues suction and reviews tomorrow, which is entirely reasonable on any given morning, and goes on being reasonable for another three weeks.

The patient, meanwhile, is not in a holding pattern. He is tethered to a drain on a ward, walking as far as the end of the bed and no further, losing appetite and muscle, and exposed to whatever is circulating that month. If he was marginal for an operation on day ten, he is further from it on day thirty, and the original judgement has quietly made itself more correct. None of this is written down as harm, because it does not arrive as an event.

What the patients actually have

I have now placed endobronchial valves for air leak in close to forty patients, some of them referred as second opinions by thoracic surgical colleagues. The striking thing on looking back over them is how little they have in common with each other.

Sixteen were on veno-venous ECMO, ten of those during COVID. The rest had pneumonia of nearly every description — influenza with staphylococcal infection, streptococcal pneumonia, tuberculosis. Some had severe COPD. Some had pulmonary fibrosis. Several were leaking after lung surgery, one after pleurodesis for mesothelioma, one after a needle biopsy of a large tumour.

Persistent air leak is not really a disease. It is the common endpoint of a dozen unrelated injuries to the lung, which is why it turns up on respiratory wards, intensive care units and surgical wards alike, and why no single team has ever quite owned it. In the ECMO patients the leak was the specific thing preventing a wean; in the others it was the thing preventing a discharge. What all of them shared was the one fact that matters for the argument: an operation was not available to any of them. That is not a description of an unusual population. It is the referral criterion.

I have been asked to talk through these cases with units around the UK and abroad, and the same question comes back every time — not whether the technique works, but who exactly it is for.

What the evidence will and will not carry

The standard's caution is not misplaced. The published literature on valves for air leak is a collection of case series, including my own, and case series are assembled from patients in whom something had to be tried. There is no randomised comparison. Anyone presenting this as established therapy is over-reading what exists, and the honest summary is that it often works quickly and that we do not know how it compares.

That is what PRO-SEAL is for. It is a randomised trial of three approaches to persistent air leak in secondary spontaneous pneumothorax, sponsored by North Bristol Trust and adopted by the NIHR, and I was involved in the protocol before I was involved in delivering it. At Guy's and St Thomas' all three options are already our standard of care for these patients, which is part of why we were able to take the study on. I am co-principal investigator at the site, and the only person delivering the valve arm.

The second clock

There is a change that would help now, costs nothing and needs no trial.

The standard starts a clock when a leak becomes prolonged. It starts nothing at the moment somebody writes not fit for surgery in the notes. That entry currently closes a conversation, and it should open one — the point at which the alternatives are named, a referral for them is made, and a date is put on reassessing rather than on continuing. Making the surgical handover measurable was the right first move. Measuring what happens to the patients the handover cannot help is the next.

What my own patients taught me about waiting

The argument above is not one I arrived at from the literature. It came from lung volume reduction surgery, where I operate on some of the most emphysematous lungs anyone operates on, and where a leak past a week used to mean what it means everywhere else — keep the drain in and hope.

On the occasions I took those patients back to theatre, I found something I did not expect. The leak was almost never at the staple line. It was somewhere else in that diseased lung entirely. Once it was found and stapled off, the leak stopped within twenty-four hours and the patient went home.

That changed how I think about the whole problem, and taking those patients back early is now my standard. A leak at the staple line is an operation healing. A leak from elsewhere in a badly damaged lung is not healing at all — there is no process underway that waiting allows to finish.

The lesson generalises past my own theatre list. In a patient with severe underlying lung disease, the leak is not going to stop on its own. The weeks spent finding that out are the weeks that cost them.

← Back to the Journal