Bilevel therapies work by alternating pressure between two levels: the IPAP(Upper-Pressure Level) and EPAP(Lower Pressure Level). Pressure support(PS) is the difference between IPAP and EPAP.

Square Bilevel Pressure Waveform

Bilevel therapies are particularly effective at treating flow limitation because of their use of this pressure support(PS). PS allows for a decrease in respiratory effort and an increase in the pressure gradient between the lungs and the PAP mask. This increased pressure gradient allows air to flow faster through a narrow airway, essentially eliminating the decrease in flow seen with flow limitation. You can think of this phenomenon as essentially simulating the properties of a large airway through a physically small one. Also, PS has minimal impact on the physical size of the airway. Only EPAP affects the physical size. EPAP is limited on the degree to which it can physically expand the airway, sometimes not even resolving flow limitation. That's where PS comes to the rescue. Keep in mind, however, in some patients EPAP is sufficient and they will be able to use CPAP effectively.

Standard BiPAP can be used to achieve a constant level of PS throughout the night. This therapy works for many patients however it does have the tendency to cause central sleep apnea(CSA) in some cases. This is due to the varying ventilatory demand between sleep stages, the varying diameter of the airway and excessive PS during arousals. This can result in under and over ventilation throughout the night. When over ventilation occurs, CSA arises due to excessive removal of CO2 reducing the respiratory drive. This is where ASV comes in. ASV only provides PS when needed, preventing the large changes in ventilation that cause CSA.

Moreover, Krakow et al demonstrated that ASV is useful for OSA patients with "Expiratory Pressure Intolerance" or OSA associated insomnia. In fact, it was found that many patients initially deemed CPAP intolerant were actually tolerant of advanced PAP therapies like ASV and Bilevel. Additionally, many of these patients treated by Dr. Krakow also had sleep-disordered breathing with flow limitation which was much more effectively treated with advanced PAP. As you can see below, ASV greatly increased the time spent with normal non-flow-limited breathing.

Percent Normal Sleep Breathing Time (%)

Additionally, this figure also demonstrates that this increase in normal breathing was also accompanied by a reduction in the number of respiratory effort related arousals (RERA). These arousals usually occur as a result of flow limitation, so it makes sense that they would be reduced.

Respiratory Effort Related Arousal (RERA) Index

ResMed vs Resperonics ASV:

There are two main bilevel ASV machines currently on the market. The ResMed Aircurve model and the Philips Respironics model. They each have completely different algorithms for manipulating pressure. Specifically, the IPAP(Upper-Pressure Level) and EPAP(Lower Pressure Level). Pressure support(PS) is the difference between IPAP and EPAP.

ASV Algorithms

The Phillips algorithm focuses on peak flow. This is essentially the peak of the airflow waveform. It is usually representative of the cross-sectional area of the airway obstruction in OSA. In OSA peak flow always decreases during flow limitation and hypopnea. When peak flow decreases the algorithm immediately increases PS and even adjusts PS within the breath based on the current flow. It adjusts PS to roughly maintain the average peak flow over the last four minutes. It also increases EPAP when obstructive events are detected.

The ResMed algorithm is based on an estimate of minute ventilation. This is essentially the area under each breath flow curve averaged over a minute. The algorithm continuously compares the current flow curve with the expected curve that must occur to achieve the goal ventilation. If there is a difference between these two curves, then PS is increased to compensate and correct this difference. It also increases EPAP when hypopnea, apneas and flow limitation. ResMed addresses flow limitation with EPAP rather than PS.

Although Krakow et al demonstrated that the ResMed algorithm was highly effective I believe the Philips algorithm may be more effective. In UARS and SDB with flow limitation, breaths are usually prolonged with a decrease in peak flow. Since flow limited breaths are prolonged in length, this can sometimes result in a constant minute volume despite the decrease in peak flow. As you can see below these intermediate breaths have similar inspiratory volume (area under the curve) to the normal ones although they are flow-limited. In cases like these, the ResMed algorithm will be unable to detect flow limitation and increase PS. Instead, using the AutoSet algorithm, it may slowly increase EPAP depnding on the degree of flow limitation. It is my opinion that, due to the pressure gradient, PS is more effective than EPAP at addressing flow limitation. Additionally the adjustment rate on the AutoSet algorithm is much too slow to treat flow limitations before arousals occur. Essentially the ResMed algorithm can only detect and correct flow limitation with PS when it is accompanied by a decrease in minute volume: this only occurs in severe inspiratory flow limitation. Conversely, the Phillips algorithm will increase PS even if a breath is 5% flow limited. It is much more sensitive.

In conclusion, I do not refute the results produced by Krakow et al, however, based on the inherent algorithms, I believe the Resperonics ASV may be more effective if studied and tested in the OSA population. Keep in mind only ResMed ASV has been studied and applied to OSA.