Test and therapeutic effectiveness ranking for 6 popular CPAP devices [ENG]

I invite you to another great CPAP test performed by our blog. I would like to remind you that earlier we tested the pressure stability and performance of selected devices. This time we decided to compare the most critical (and also the most difficult to evaluate) element, which is the algorithm. We invite you to read this article.

The questions we tried to find answers to:

  • Which of the tested devices was the most effective in dealing with apnea treatment?
  • What are the differences between the devices? How big they are?
  • Does the price of a CPAP correlate with its effectiveness?
  • Does the indication of the AHI displayed by the machine correlate with the actual result?
  • Do we need to titrate with automated machines and can we titrate with an automatic machine?
  • How to use this knowledge when buying CPAP for yourself?


For the test we chose the most popular CPAP automatic devices available on the Polish market and those which in our opinion were worth verification. These are in alphabetical order
  • APEX iCH Auto
  • DeVilbiss Blue Auto Plus
  • Löwenstein Prisma 20a
  • Philips Dreamstation Auto CPAP
  • ResMed AirSense 10 AutoSet
  • ResVent iBreeze Auto CPAP


The test assumed control of parameters describing the effectiveness of obstructive sleep apnoea treatment with the use of polysomnography machine. Each device had one night's sleep and as a control one without CPAP. The order was random.

We made sure that the results were as reliable and repeatable as possible. Our "test subject" was always healthy and rested;  no drugs and stimulants that could affect the result. The length of sleep was similar, sleeping always in the same place, with minimal disturbance to the test.

The same mask and identical parameters have always been used:

  • pressure 5-20 cm H20
  • comfortable moisturizing (Auto)
  • exhaust relief (EPR) at maximum (3)
  • the standard algorithm, no other functions
We intentionally did not set the pressure parameters within the optimum range to see how much the automat can handle the setting of the correct pressure on its own. For this reason, the AHI results obtained during the test will be worse than during the actual therapy with the device set up correctly.

We evaluated the following parameters:

  • AHI value and pressure of 95% from the examined night determined by the device
  • subjective assessment of sleep quality/respiratory comfort
  • objective result of the therapy determined by the PSG, including AHI, ODI, hypopnea index, sleep quality assessment


The polysomnographic system used in our study was NOX A1. It is a widely recognized and proven device on the market that allows for PSG measurement at home (including EEG channels allowing for precise determination of sleep phase).

In the study we recorded the following channels:

  • EEG: C3-M2, C4-M1, F3-M2, F4-M1, O1-M2, 02-M1, E1-M2, E2-M1 
  • EKG
  • RIP: abdominal and thoracic
  • Mask pressure/ nasal cavity pressure
  • Volume/light
  • Pulse rate and blood saturation
Evaluation of results using Noxturnal app.


At the first moment, we relied on automatic analysis, given that it would be the most objective. It turned out, however, that depending on the algorithm used, we obtain different results, so the final decision was made to analyze the results manually, by a high-class professional. The analysis was conducted in a "blind" way, i.e. the doctor did not know which results were available from which CPAP device, which was supposed to help to maintain the objectivity of the assessment.

In the physician's opinion, we made a good choice of the patient. The case was assessed as quite typical. The quality and course of sleep were very similar at subsequent nights, with clearly marked phases. There were no objections to the quality of the data, so the evaluation of the results was not difficult.


Male, 43 years old, obese. With a metabolic syndrome, history of cardiac problems in the past. Does not smoke cigarettes, does not take drugs that may affect apnea. No coexisting pulmonary diseases or anatomical defects obstructing breathing. OSA diagnosed and successfully treated with PAP therapy for about 2 years. Therapeutic pressure of EPAP, tolerating obstruction at the level of 7.5-8.0 cm H20 (determined method of manual titration).

During the control night, which was part of the study, obstructive apnea was severely observed. AHI index 80.2, of which hypopnea was 39.9. ODI 63.9. Snoring 43.6% of sleep. Paradoxical respiration visible in the record.


Although we have taken great care to ensure that our research is as reliable and objective as possible, it should be remembered that this is not a scientific research. In particular, it should be noted that the results may vary from person to person and that only one night for each device was tested. We may have unconsciously omitted an external factor that affects the results.

On request, I will make the "raw data" from the individual studies available to people who would like to undertake the analysis themselves (of course we are talking about professionals).


Average therapy results based on the AHI index determined by the CPAP device itself are 2.4 (median 1.3), which is a fantastic result considering that we started with 80.2 and that the automats were not previously set on the basis of the titration result. Of course, assuming these values are true.

Now get ready for a bucket of cold water. The actual results of the polysomnographic treatment are 13.35 (median 11.90). The best device was able to lower the AHI to 2.7 (ResMed), and the worst results are 29 (Resvent)! Only two devices were able to lower AHI below 5 (ResMed and DeVilbiss), which, as I recall, is the conventional limit for the diagnosis of apnea.
/the line is the value declared by the unit, the bars are the actual result/

It is noteworthy that basically all devices have coped well with apnea, but not all devices have managed to cope with hypopneas, which is the main component of AHI in every case.

Apart from the AHI value, the indicators that clearly differentiate the effectiveness of the therapy is the ODI index (average 20.45, minimum 9.8 Philips, maximum 38.6 APEX). 

and % of snoring ( average 21%, minimum 0.3 ResMed, maximum 58.5 APEX).

I would also note the arousal indicator as an important component of "compliance". (average 6.97, minimum 3.1 Lovenstein, maximum 16.2 DeVilbiss). However, the physician pointed out to me that this indicator can be easily disturbed. 

I am also interested in the analysis of the pressures of 95% achieved by devices:

With a blue rectangle I marked the lowest effective therapeutic pressure determined by the manual titration method. It is clear that only Prisma "shot" into the range, and the vast majority used higher pressures. AirSense stands out here because, despite its relatively low blood pressure, it was able to effectively reduce snoring - so from a pressure/efficiency point of view it was the best. By titrating with a Philips machine we will achieve a lot higher than the required pressures.

This is the dependence of the pressure on the effectiveness of the therapy (one should be careful with interpretation, however, as the pressure is not the only important indicator of the therapy):


Below is a discussion of the results of the individual units, in order of order of the best in my opinion:

IST PLACE ResMed AirSense 10 AutoSet

AHI device: 1,3
AHI real: 2,7
Pressure 95%: 14,38

The undisputed winner of our test. Although I will honestly say that I had some concerns about the results, because the beginning of the night, was not too comfortable for me because of too low pressure.

AirSense has lowered AHI most effectively in all devices and is best at reducing snoring. Good ODI. Interestingly, during the night I had a fairly high pulse rate of 70 and a low average saturation of 92.8%. The cause was UARS appearing during the REM phase.

II PLACE Philips Dreamstation Auto CPAP

AHI device: 2,1
AHI real: 5,9
Pressure 95%: 16,7

The deserved second place, although I have to make a disclaimer that the night recording on this CPAP was the worst. For some reason I had a problem with falling asleep and the dream was " shredded". So maybe the result could have been even better if you had repeated the test. Interestingly, despite these problems I found the sleep comfort better than in ResMed.

Dreamstation did its best to reduce the ODI and eliminate snoring in an exemplary way. I noted the lowest pulse rate of 59 and the best saturation of 94.1%. 

III PLACE DeVilbiss Blue Auto Plus

AHI device: 7
AHI real: 4,3
Pressure 95%: 15

Very positive surprise of our test. Blue was the only one of all the devices to overestimate AHI results. The subjective sleeping comfort was low, but our physician pointed to the model algorithm's work in terms of quick and decisive response even to small disturbances that were omitted by other devices. Thanks to this, it effectively eliminated the hypopnea.

Unfortunately, DeVilbiss is not able to cope with snoring completely and his rate remained at 30.8. This resulted in a high arousal rate of 16.2.

Lovenstein Prisma 20a

AHI device: 1
AHI real: 12,7
Pressure 95%: 12,5

Prisma managed to manage in our test quite decently, despite the relatively high AHI index. It deserves additional points for high comfort from the lowest pressure (probably thanks to the most efficient pump in the test). I had the lowest indicator of awakenings on it. It did quite well with snoring (7.2) and ODI (13.9). The basic algorithm did not do the best for the hyponeas.


AHI device: 0,5
AHI real: 29,5
Pressure 95%: 11,6

The results of this machine I was probably the most curious. The Chinese made a beautifully looking and well made machine that you might like. Unfortunately, the algorithm itself is much worse. The device completely failed to cope with my apnea in the back and REM position. It did not eliminate snoring. He is completely unable to cope with the hypneas I had a very high ODI (27).  What's interesting, however, I had quite high comfort at the start thanks to the IPAP raised by as much as 4.5 cm H2O. I don't know if such a thing is "legal" but I didn't complain at all. The device has some problem with the relief of exhalation, which does not synchronize well with the breath. Unfortunately, after this test I can't recommend the purchase of this device.


AHI device:?
AHI real: 16,9
Pressure 95%:?

It might seem that AHI is not at all much higher than Prisma and is clearly better than iBreeze-why does it give the device the last place?

Sleeping with this device was much less comfortable for me than sleeping without the device. Respiratory resistance is absurdly high. The pressures generated are too low. I snored even more on it than without the CPAP. Number of desaturation and therefore the highest of all ODI devices (38.6).

In my opinion, this device is not suitable for human use and should be withdrawn from the market. Do not buy it, because you will only be discouraged from CPAP therapy.

Summary on the table:


If you would like to read the detailed data, please read the summary table:


Let us begin by answering the questions that were asked in the beginning:

  • Which of the tested instruments was the most effective in dealing with apnea treatment?
Our battle was won by ResMed AirSense AutoSet, only slightly weaker was Philips Dreamstation Auto. As you can see, it's no coincidence that both of these devices are very popular among physicians and patients. If you still had doubts whether companies sometimes "cut off coupons" from a well-developed brand, I think that our test should dispel these doubts. It is clear that experience in this industry pays off. At the other end of the ranking were Chinese devices. To be honest, APEX did not surprise me, but I am very disappointed with RESVENT, because it seemed that their newest device could already compete with the best. Unfortunately, the device looks better than it works.

  • What are the differences between the devices? How big they are?
    As can be seen, the differences are huge already in the basic scope, i.e. lowering the AHI. The difference between the best and the worst AHI is almost 30 points, i.e. as much as it separates a completely healthy person from a severe apnea patient.
      I am surprised that many devices do not cope with snoring at all. Clearly, we can also see the problems of these cheaper devices with hypopnea. Probably if you were testing a person who only has apnea, then the results of these weaker instruments would be much better. The problem with hypopnea in my case was reflected in high ODI.

      • Does the price of a CPAP correlate with its effectiveness?

      I would like to say that it is different... but unfortunately the high quality has to be paid for. Our comparison shows that more expensive and branded devices are clearly better than their cheaper Chinese competitors. Devices in the middle price range are also in the middle when it comes to effectiveness.

      • Does the indication of the AHI displayed by the machine correlate with the actual result?

      I think it's gonna surprise a lot of people. Indication of AHI from the instrument in most cases has nothing to do with its actual value. Therefore, CPAP should not be treated as a diagnostic device, and the data presented on it should be treated only as an approximate value. This approximation is close to the truth only in ResMed, Philips and DeVilbiss.

      • Do we need to titrate with automated machines and can we titrate with an automatic machine?
      Although I will probably not change the world in this area, I hope to give at least some thought to it. Official recommendations that the titration should be on a manual device, under the direct control of the physician do not come from nothing. The machine can allow a therapeutically effective pressure but it will almost never be the optimal value and usually overestimated. The results of the X auto machine titration may be wrong for the Y machine therapy.

      • How to use this knowledge when buying CPAP for yourself?
      The main conclusions are as follows:

      - the automatic cpap is very different from each other
      - if you have a severe case of apnea, with hypopnea and snoring, it is best to reach for a branded device
      - if you have such a possibility, it is worthwhile to verify the effectiveness of the therapy with the PSG device.
      - good AHI indications on the device do not guarantee high effectiveness of the therapy at all
      - so far, it is still worth avoiding the purchase of Chinese devices - their algorithms have not passed our test successfully
      -when you buy CPAP, remember that you need a good titration and setting of the instrument.


      I'm planning further texts on the basis of the data gathered in this trial. A very interesting thing that was noticed by the doctor who consulted the results, are completely different respiratory patterns generated by different machines. I will also want to delve deeper into the analysis of the ODI indicator and deepen the analysis by a paradoxical breath. I wonder how to set the optimal PAP therapy for myself with all this knowledge. In order to do so, I will undergo in-depth diagnostics and inform you of what we have been able to determine.

      If there is a lot of interest in the study, I do not rule out repeating the study on a wider sample, as well as the BiPAP device test. In the course of this test, we also tested 2 BiPAP instruments, but the data collected are not reliable due to the wrong setting of the therapy parameters. In the case of BiPAP, it is a mistake to say that the device has no preset pressures and that is why our test was not successful.

      If there are any other issues worth an in-depth analysis, please write to me.

      Brak komentarzy:

      Obsługiwane przez usługę Blogger.