An interview with Dr. Michael Berthon-Jones

Dr. Michael Berthon-Jones has been closely involved in the field of respiratory medicine and the development of CPAP devices for most of his working career.
After completing his medical degree at Sydney University, Dr Berthon-Jones spent two years working with Dr Russel Vandenberg at Royal North Shore Hospital, in the area of respiratory medicine.
He was working on a PhD thesis studying the control of breathing in sleep, both in normal subjects and in obstructive sleep apnea, when his supervisor, Professor Colin Sullivan, first proposed using nasal CPAP to treat obstructive sleep apnea. Dr Berthon-Jones worked on the first take-home CPAP devices and the first take-home customized CPAP masks.
Later he moved on to carry out research on animal models of sudden infant death syndrome, then spent several years teaching respiratory physiology, before moving into the development of software for respiratory function laboratories and sleep laboratories.
In 1990, he joined the newly established ResMed company where he is currently Chief Scientific Officer.
Here Dr Berthon-Jones speaks about the work behind the development of CPAP devices.
What inspired the concept of an ‘‘intelligent’’ automatic clinical CPAP device?
It was Professor Colin Sullivan who came up with the original concept of an intelligent CPAP device. He was working with patients who were unable to get to sleep on high levels of CPAP. He felt the problem could be resolved by a device that would keep the pressure low until the patient was asleep, and then quickly put the pressure up until
he was well treated.
That is where the idea of a CPAP machine that would increase pressure if there was snoring, was born. I was asked to help study how well it worked at Colin’s sleep laboratory at Royal Prince Alfred Hospital and to suggest improvements.
Later we realised that an intelligent CPAP device has perhaps three uses. Firstly it can keep the pressure low while you are waiting to go to sleep. Secondly, it can find the right pressure for you without having to wait months, sometimes many many months, for a night in a sleep laboratory where your pressure can be titrated. Thirdly, by
only giving you high pressure when you need it, you get less pressure-related side effects, such as mouth leak and blocked nose. Some people only need high pressures when they are on their back, or dreaming, or if they have a stuffy nose or a cold coming on.
What sort of problems did you encounter in developing it?
The biggest problem was getting a comfortably sealing mask. In those days we used to make custom individualized masks, at great expense. And we’d glue the mask on with a silicone adhesive. This was a very messy, time consuming and expensive business, but it produced a superb seal.
What led you to think about developing an automatic CPAP device for home use?
That was really the goal from the outset. However, it was much easier to come up with a device that worked in the hospital only, because we didn’t have to worry about cost, or how big it was, or how ugly it looked. We could concentrate on making it work superbly. We ended up with AutoSet Clinical, of which I am still very fond. It was too big and expensive to go home with a patient, but it did a very good job at automatically determining how much pressure the patient needed. They could then go home on a conventional CPAP machine set at that pressure. Over time, we came up with AutoSet T, the take home device, which is much smaller and simpler for a patient to use.
What led you to incorporate the idea of flow limitation into the algorithm, years ahead of its time?
The initial device (based on responding to snoring alone) could only respond to reasonably loud snoring. This is because it had to ‘hear’ the snoring through the roar of the motor and the mask. That sort of approach could get the pressure reasonably close reasonably quickly, but it turned out that it was generally insufficient. We now know why. If
the pressure is just enough to eliminate snoring, the airway will still narrow to the point where it is very difficult to breathe. This is called silent inspiratory airflow limitation.
While I was doing my PhD, I noticed that normal subjects, while they were asleep, showed a characteristic flattening of the inspiratory flow-time curve, which went away the instant they awoke from sleep. I noticed that if you put obstructive sleep apnea patients on CPAP, and then lowered the pressure a bit so that it was enough to prevent
snoring, they would also show this flattening of the inspiratory flow-time curve. Indeed, if you put them on just adequate CPAP and pushed them a bit harder by giving them some carbon dioxide to breathe, this would also produce this characteristic flow limitation, so that even though they tried harder and harder to breathe, they didn’t get any more air for their efforts.
So I thought here was the way to ‘fine tune’ an automatic CPAP machine: increase the pressure if you see apneas, or if you hear snoring, but increase it that last little bit if you see this characteristic flattening of the flow time curve.
Why is it important for an automatic CPAPdevice to respond to flow limitation, snore and apnea?
The characteristic flattening of the flow-time curve caused by flow limitation is the very best signal for fine-tuning the pressure, once you have eliminated apneas and snoring. But if you are just falling asleep, you can go very quickly from having a totally open airway to snoring very loudly, in a way that produces somewhat chaotic or messy flowtime
curves, without seeing the characteristic flattening. So the best approach is to respond very quickly to loud snoring, and then fine tune using flattening. Rarely, you can go straight from awake and unobstructed to asleep and apneic, and so it can be useful to increase pressure in response to apnea as well. However, actual apnea is pretty rare
on AutoSet, because in most cases the responses to snoring and flattening get the pressure up quickly enough to prevent apneas.
Why doesn’t ResMed's AutoSet respond to hypopnoea?
When you are lying quietly awake, or when you first go to sleep, or when you are dreaming, you can have hypopneas (reductions in the depth of breathing) which are nothing to do with the state of the airway. For example if you sigh, which you do every few minutes, you usually have a hypopnea immediately afterwards. This can also happen if you have just rolled over and are getting settled, or if you are dreaming. And the annoying thing is that
when you are on CPAP, this tendency to have what are called central hypopneas - hypopneas that are nothing to do with the state of the airway - is increased. If you make an automatic CPAP device that responds to hypopneas, you will put the pressure up to the maximum while the patient is awake.
Do you think there is a misconception clinically that all hypopneas should be treated ?
For simple obstructive sleep apnea, central hypopneas should not be treated. They are not a disease. Everyone has them. And they don’t go away with CPAP. There is a rare and important exception: central hypopneas due to heart disease. This is called Cheyne-Stokes breathing. CPAP does help with that.
Why doesn’t ResMed's AutoSet respond to apnea above 10 cmH2O in pressure?
I mentioned before that the higher the pressure, the more central hypopneas you will have. At a pressure somewhere around 10 cmH2O, the central hypopneas become central apneas. On the other hand, the vast majority of obstructive apneas are already well controlled by 10 cmH2O, and we are only fine tuning using snoring and flattening. So it is a pretty good bet that if the pressure is already above 10 cmH2O, any apneas are most likely
central, and you should leave them alone (except in patients with central apneas due to heart failure). But if the pressure is below 10 cmH2O, most apneas will be obstructive and you should put the pressure up. There’s nothing magical about 10 cmH2O, it’s just a good place to put the line in the sand.
Can you over-treat apnea?
You can’t over-treat obstructive apnea. You really don’t want the patient having unresolved obstructive apneas. And we want not just to prevent apnea - we also want to keep the airway sufficiently open for the subject to breathe easily and regularly and stay asleep.
But you can use too much pressure. The higher the pressure, the greater the side effects. Although this has never been proven, it is rather obvious - no pressure, no side effects! So you want to use the lowest pressure possible while keeping the airway nicely open.
Likewise can a device that responds to hypopnea over-treat it ?
The funny thing is that it can both over-treat and under-treat. It will put the pressure up through the roof in some subjects, who have lots of central hypopneas. And it can completely miss repetitive severe silent inspiratory flow limitation that is totally disturbing the patient’s sleep without there being any hypopneas. If this occurs without CPAP, it is called upper airway resistance syndrome. It is just as bad for you as obstructive sleep apnea. But a
CPAP machine that responds only to hypopneas will treat your obstructive sleep apnea, and give you upper airway resistance syndrome instead.
How can Automatic CPAP devices help optimise treatment ?
CPAP devices, whether automatic or not, can tell us - the clinician, the technician - about what is going on when we are not there. Is the patient using the device? Is there a leak, and if so, when and how much? If it is an automatic device, what is the pressure doing? How well is the patient breathing? How steadily, how much? This might be particularly important if the patient also has heart disease or lung disease, or has had a stroke, and has other
reasons, apart from sleep apnea, for having abnormal breathing during sleep.

From: ResMedica - Clinical Newsletter - First Edition