This is a written version of a talk requested for PACSA 2019 in Johannesburg. It covers the horror show that is button battery ingestion and the tracheo-oesophageal fistulae that can result. The references are down the bottom on this too.
The future is a scary place. Climate change is real. The politicians dealing with it are probably still pretty dense. And the robots are coming to get us with gymnastics. Deep breaths.
Meanwhile kids will keep finding ways to provide challenges and technology is likely to help on the mission. Which brings us to the central topic of this whole thing – the little batteries that cause big havoc.
Why so serious?
Batteries aren’t new of course. Button batteries aren’t new either. The National Capital Poisons Center (NCPC) keeps track of events with button batteries and they well and truly record deaths way back to 1977.
Yet there is a clear sense that button batteries are turning up as a pretty scary problem more and more. They certainly stick in the mind of anaesthetists who’ve been involved with a case where things got serious. Even in the room at the conference there are people who can attest to an alphabet of button batteries looking like
A – Airway
B – Breathing
C – Circulation
D – Drugs
E – ECMO
Anyway, over 1990-2009 the number of children presenting to emergency departments for battery-related injuries nearly doubled. There are a number of potential explanations for this. For starters everything we use keeps shrinking and we keep buying stuff that needs batteries. There are just more things containing these discs of destruction. The US alone sees reports of over 3500 incidents of ingestion each year and that is most likely to be an under report.
The key thing that has changed with time though is he severity of injuries. Back in 1992 the numbers looked like a rate of major morbidity of 0.08% with no fatalities. What hasn’t changed since then is the rate of ingestions. The severity is a different story.
The most recent data from the NCPC suggests an incidence of major morbidity or mortality of 0.66%. That’s a sevenfold increase. So while the majority of ingestions results in, well nothing, the likelihood we’ll see one of the bad ones seems to be going up.
So apart from there just being more stuff, what else has changed?
Well in 2006 there was a big spike in morbidity. This correlates with the introduction of the 20 mm 3-volt lithium button battery to the market. Larger and more powerful, they are more likely to get stuck and more damaging when they do. In kids under 6 years of age the rate of major morbidity or death after ingestion of a button battery > 20 mm in diameter is as high as 12.6%.
One of the other key challenges with getting on top of these cases of course is the nature of the presentation. It’s not always obvious. First there’s the delay in presenting with anything. Not many kids seem to be getting watched when they swallow that battery.
Then when kids do present they forget to turn up with a shirt printed with the problem. Symptoms and signs are pretty non-specific things. Dysphagia, fever and cough are probably most prominent with some additional slight differences with different ages. Infants may just present with irritability (never happened in a baby before), anorexia and melaena. Toddlers are more likely to turn up with drooling, vomiting and dyspnea. Older kids might actually mention thoracic or abdominal pain.
And of course in all that non-specific time, tissue is getting liquefied and things can get bad really quickly. It turns out having things stuck causing necrosis in areas that contain airway, gut and big blood vessels in close proximity is not a guarantee for restful nights for everyone.
So faced with a situation of:
- Thing I will probably not see regularly (hopefully);
- Really need to do well if I do,
How do I get to be an expert? Really that’s the goal isn’t it?
Perhaps the first step is to develop an understanding of what an expert is and what they do in particular.
The expert generally has an advantage where they can utilise pattern recognition to make rapid changes in approach. They are able to get more done subconsciously because they are able to rely on what they’ve seen before to rapidly get to an answer and respond without exercising particular thought. This leaves more bandwidth for simultaneously taking information in and then acting. The type of exposure does sort of matter though. Not just for the content of what has been seen before but also how available it is to their cognitive process. And that availability is heavily influenced by how memorable that last time was.
There is a downside for the expert though. When something doesn’t match an ‘illness script’ they already recognise, they don’t necessarily do so well taking a step back to analyse and work from first principles. So that’s something we could consciously plan for to make the response better.
So perhaps if we haven’t seen things before, we should work on building up some patterns to recognise.
The First Bit
The first thing is to know what specific questions in the acute phase might help. Create a pattern of knowing the acute priorities to minimise and assess risk. So it’s worth asking (and I’m basing this on the notion that they’ve already identified it as a button battery):
- What about the honey?
- Where is it stuck?
- Are there any heralds of bad things?
Honey is actually a recommended option for trying to limit damage (as is a dilute solution of sucralfate). The theory is that honey coats the battery and prevents local generation of hydroxide ions which hopefully means less of an alkali burn. It slows down damage but doesn’t prevent damage.
What’s the evidence? Well it involves pigs. In vitro studies of various liquids in cadaveric porcine oesophagus and in vivo study of the effects of honey adn sucralfate compared to saline irrigations of batteries placed in the oesophagus of live piglets.
The National Capital Poisons Centre recommends 10 mL every 10 minutes for up to 6 doses. The suggestion is that you can try this up to 12 hours but of course this doesn’t change the fact that you should be trying to progress things quickly. The goal is to get that battery out within 2 hours of presentation wherever possible.
Where is it stuck?
This is the really key issue. Patients need an X-ray. Probably a while back. The goal is to figure out where the button battery is and where the highest risk of injury lies, particularly where it comes to surrounding tissues.
There are 3 key points where things get narrow and that pesky little localised diathermy device of a button battery is likely to get stuck. The commonest site is the thoracic inlet in the upper third of the oesophagus where it gets held up at the cricopharyngeus muscle. For those wanting a refresher on where that lies you’ll be looking between the clavicles.
The second key spot is the is in the mid-oesophagus which corresponds roughly with the left main bronchus and the aortic arch. Obviously that’s a worry because erosion in that area comes with incursion on very big blood tubes.
The final one of course is at the lower oesophageal sphincter. While the smaller button batteries may well clear all of those spots, anything bigger than 20 mm is enough to make you sweat.
There’s another key thing to look at on the X-ray. The button battery itself will give a double halo appearance on X-ray, at least when seen face on. If you look from the side of the battery you can see the step off and that corresponds with the negative pole of the battery. This is where the damage is going to be greatest. Wherever possible understanding where that negative pole can help with appreciation of risk to surrounding tissues.
Heralds of bad things
This just means bleeding. Brumbaugh et al showed that up to 70% of fatal cases of haemorrhage present with some history of bleeding. It doesn’t need to be big bleeding. Any bleeding should tell us ‘there is something happening here that is joining a tube which isn’t supposed to have blood in it, the swallowing one, and a tube that is supposed to be full of blood’.
This can best be described as *not desirable*.
The Priorities of Treatment
Time is everything.
The recommendation is to try and get that battery out within 2 hours of presentation if it’s stuck in the oesophagus. That doesn’t really depend on when they present of course. If they present at minute 10 after they ingested the thing, you have 2 hours. If they present 4 days later, you still want it out really quickly. It’s just the risk profile that is a bit more terrifying.
Now there are some patients who don’t need an operation immediately. This would be kids > 5 with a battery in their stomach that is less than 20 mm in size. You might be able to watch those ones a little but repeat imaging is probably a good idea.
But every patient where the battery is in the oesophagus needs that power cell out. Every under 5 year old patient with it in the stomach but with a battery that’s over 20 mm needs it out too. It’s not quite with the same urgency.
The Right Place
Of course what we’d really like is also have the patient in the right place. And while you might think the right place is ‘anywhere I am not’, we’re talking about risk assessment. The high risk patient should be in a place with the capability to get onto ECMO or cardiopulmonary bypass quickly in the ideal world.
The ideal world.
It is fair to say that the ideal world isn’t always where we’re living when a kid pitches up with a late presentation and a bit of bleeding. For the practitioner out there in the periphery the real challenge becomes dealing with that balancing act:
- Transfer the patient so cardiothoracic surgery is an option but that could take ages.
- Hold onto the patient to get the thing out quickly.
And the only people who can make that call are the clinicians there at the time.
Yeah I sort of wimped out there.
At the very least if we’re in a high risk situation the team make-up should reflect that. You need someone to get the battery out with the clever camera thing and if it’s a lower risk case without sentinel bleeds that might be enough. You do probably need to at least consider if there’s a role for ENT. You really want the cardiac back-up when it’s high risk.
And sometimes a friend. An anaesthetist friend.
And another friend – a big cannula (probably with some invasive monitoring tagging along).
The 4 Pattern Groups
To get back to that expert mindset we can actually just tap into patterns that we’ve worked on for years. There is:
- The unfasted patient (because they are).
- The patient who will have a big bleed (because they might).
- The patient with an air leak (because that’s possible and painful for all involved).
- The patient with a foreign body (because we’ve all met a kid with a misplaced peanut).
The first one is relatively easy to account for – is your highest risk the ‘they’re not fasting’ issue? In that case maybe rapid sequence intubation should be your plan. Or is your main worry the airway? If it’s the latter then we’re starting to get into TOF territory. More on that in a tick.
With all of these of course we need to be able to change direction quickly. This means an ability to respond when things go bad and a plan for what to keep an eye on.
If our worry is bleeding then invasive monitoring should be enough to let us keep up to date and respond and big IV access with blood products available has to be part of our plan.
If our concern is the potential for airway fistulae to be created then we probably need to be able to diagnose an airway fistula. Which means it makes sense to have a flexible bronchoscope there. More importantly we might want to keep an eye on the thing that would tell us if a new join popped up – end-tidal CO2.
There’s one case report out there of a tracheo-oesophageal fistula declaring itself because the end-tidal CO2 suddenly jumped up with the gastroscope CO2 insufflation. If the expert thinkers challenge is being able to step back and analyse when things need a non-programmed response, anticipating in advance what might happen and patterns to keep an eye for can bridge that gap.
The Simple Plan
So after all that we have a really simple set-up:
- Choose the right place.
- Summon the right team.
- Prepare the patient and their family (this is actually a huge topic but we’ve still got some TOF work to do).
- Place a cannula first.
- Choose your induction that accommodates what you think is the priority.
- Set-up invasive monitoring.
- Be ready for mayhem, with options to change directions.
Well there’s a couple of other things to think about quickly:
- There’s a pretty good chance that a patient with this sort of battery ingestion they haven’t been keeping with their intake. That might make itself very clear once you’re under way. A fluid load to kick off might well have a role.
- What about airway oedema? It’s pretty hard to know exactly what that airway might be like so we can at least think airway oedema and have some smaller endotracheal tubes available.
- Oh and just a note about those high risk cases. A central line might be a really useful thing. Given that you may not quite know what’s going on up in that neck, maybe make it femoral.
- And at the end you might want to think about whether time is needed for healing or if the breathing snorkel should come out straight away. You don’t even have to do it alone you can do that as a team. In fact you probably should do it as a team.
Once that thing is out (yep, I let us fast forward) the oesophagus and airway should be examined for evidence of the severity of injury and then admitted to the right unit (like even an intensive care area initially) with an eye being kept on new respiratory or haemodynamic signs. That would suggest an evolving injury (or just one that’s obvious later). Nutrition needs to start (either via NG placed delicately at the time of removal or via another means). Serial evaluations will be part of reality and that might require further endoscopies as well as CT or other options to see the things below the mucosa.
But Wait, There’s TOFs to Deal With
Oh, yeah. I was asked about TOF when they first brought this up.
Well the simple thing to say here is no one really knows what they’re doing.
Wait, that’s a little bit the wrong way.
What I mean is there isn’t a single unified plan for patients with TOF and that suggests that no one can guarantee a single option is the best forward. Through my training I was always taught to keep the patient breathing. It keeps you oxygenating, stops you inflating the stomach and maximises your chance of success. And clever people told me that. So I just listened.
If you go to the literature though, it turns out other people must be saying the literal opposite. Knottenbelt et al described practice across Australia and New Zealand back in 2012. They looked at 101 patients and 81 of those were intubated in theatres. 32% had an IV induction and only 68% had inhalational anaesthesia as the chosen plan. At induction 21 patients received suxamethonium, 8 (that’s 10%) had a non-depolarising agent and the remaining 52 (that’s 64%) were kept breathing. 43 of the 106 total patients had a bronchoscopy before the surgeons really got rolling.
Fast forward to 2019 and Van Hoorn et al have reported on 101 consecutive patients (101 is apparently the magic TOF number). 86 patients were intubated in theatres. 52 of those ptients had IV induction and only 30 had sevoflurane (4 were unknown). 74 of the patients had non-depolarising muscle relaxation and 4 had suxamethonium (8 are unreported and you can’t actually tell when they received their muscle relaxation anyway). This time 42 of the patients had bronchoscopy prior to the surgical action.
So what are we to make of this?
Well I guess that there is more than one way to do most things. Things I had always thought were frowned upon seem to be the go to option for some people. What I would say though is that once you’ve given the muscle relaxant you’re pretty committed in one direction. If you can get by with low pressures and maintain oxygenation while spontaneously breathing you should hopefully avoid developing gastric insufflation.
A couple of notes though:
- Sometimes you will still end up providing muscle relaxant after assessing that spontaneous breathing is just not cutting the mustard (sidenote – mustard seems like a pretty weird thing to bother trying to cut so this makes no sense to me).
- The old anatomical classification of trachea-oesophageal fistula doesn’t necessarily apply here because the button battery isn’t really an embryological agent. This only argues more for the value of using the flexi bronch.
So as there’s no real guidance as to a single technique and no particular evidence of definitely superior safety or success with one technique over another for a patient with trachea-oesophageal fistula, I can only report what I do.
- Inhalational induction with spontaneous ventilation.
- If we’re there to correct the TOF, then there will be cutting. I’d rather avoid opioids though so a combination of blocks (ESP/serratus anterior plane/intercostals), ketamine and dexmedetomidine (noting that in our button battery situation that might be tempered by that whole catastrophic blood loss thing).
- A flexible bronchoscope available for me thanks. Being able to examine the airway when there’s a chance that the airway is stuffed seems like a sensible option to have available.
- Paralyse once that annoying join is sorted.
In the meantime we should probably focus on the complications bit. In the van Hoorn study they made a good effort to look at changes in oxygen and carbon dioxide:
- Hypoxaemia happened in 75 of the 101 patients and 28 of those were severe (meaning < 80% saturations. 2 dropped below 50%.
- 28 had hypocapnoea and 46 had hypercapnoea. But that also meant he pH was 7.2 or below in 33 of the patients and < 7.0 in 3.
Unfortunately it’s impossible to say which techniques were associated with which complications so there’s still nothing to be definite about.
So the future is a horror show. And sometimes that future will involve very scary cases with very small batteries.
We won’t see them often so we have to do our best to employ the things we already know to guide us to address the right priorities and change streams rapidly when things go south.
We should control what we can by minimising damage, moving quickly and getting the patient to the right place with the right team.
We should use the strategies we know in the situation that applies – the unfasted patient, the patient who bleeds, the airway that leaks.
And we should be looking for the signs that tell us to switch streams.
Then we should breathe.
That one great review on button battery ingestion is this one:
There’s a more recent one which actually reads as very similar but has some good images:
The guideline from NCPC is here.
Now I mentioned experts and how they think. There are two things worth a look that underpinned some of the stuff on that. The first is a paper:
Plus the editorial and review that comes with it (with Dr Majorie Stiegler as one of the authors again):
That case report about capnography and picking up a tracheo-oesophageal fistula is this one:
Garcia Getting RE, Harris CL. Sudden Increase in EtCO2 During Upper Endoscopy Under General Enotracheal Anesthesia Suggests the Presence of Tracheoesophageal Fistula: A Case Report. A & A Case Report. 2017;9:109-11.
The first of those TOF papers would be this one:
And the more recent one is this one:
van Hoorn CE, Costerus SA, Lau J, et al. Preoperative management of oesophageal atresia and tracheoesophageal fistula: An analysis of data of 101 consecutive patients. Pediatr. Anesth. 2019;29:1024-32.