CPAPtalk.com

This is good work. Thanks Velbor.

I gotta go continue "soccer day", but from what I've read so far it's obvious you've put a lot of work into this. It bears further reading that is for sure!

You are basically saying that the only problems with mouthleaks is the fact they dry your mouth, and cause arousals. According to your understanding they have no impact on the machines' ability to identify apnea, hypopnyea or flow limitations, and therefore have no effect on the machines' abililty to give proper treatment.

Velbor wrote:Nor have I seen documentation that during an apneic event, persons reflexively attempt to open their mouths.

Mouth Breathing Compromises Adherence to Nasal Continuous Positive Airway Pressure Therapy

No distinction was made between mouth breathing and mouth opening at the end of apnea caused by choking.

Baseline Demographics and Sleep Characteristics
We studied 51 patients (4 women). MBs and NBs did not differ with respect to age, sex, body mass index (BMI), neck circumference (NC), ESS, AUDIT score, TST, TST while supine, rapid eye movement (REM) sleep, wakefulness after sleep onset (WASO), sleep efficiency, RDI while supine, or pulse oxygen saturation (Spo2) awake. Meanwhile, total arousal, respiratory arousal, RDI, and an oxygen desaturation index of 4% (ODI4) were higher in MBs, but the amount of delta sleep was lower

.

Effect of nasal or oral breathing route on upper airway resistance during sleep

During wakefulness, upper airway resistance was similar between the oral and nasal breathing routes. However, during sleep (supine, stage two) upper airway resistance was much higher while breathing orally (median 12.4 cmH2O·L–1·s–1, range 4.5–40.2) than nasally (5.2 cmH2O·L–1·s–1, 1.7–10.8 ) . In addition, obstructive (but not central) apnoeas and hypopnoeas were profoundly more frequent when breathing orally (apnoea-hypopnoea index 43±6) than nasally (1.5±0.5).

Obstructive sleep apnoea and oral breathing in patients free of nasal obstruction

In conclusion, apnoeics spent more time breathing orally and oro-nasally than simple snorers, and the apnoea/hypopnoea index is a major determinant of the time spent breathing orally and oro-nasally.

Mouth Breathing in Obstructive Sleep Apnea prior to and during Nasal Continuous Positive Airway Pressure

The lowest SaO2 measured was 82.9 ± 4.7%, and increased to 87.5 ± 2.7% under nCPAP. The mean nCPAP was 7.8 ± 1.6 cm H2O. MB event index per hour of sleep decreased from 35.2 ± 19.7 prior to treatment to 5.0 ± 5.2 under nCPAP (p < 0.01). In 52.2 ± 27.4% of obstructive respiratory events, MB started at the end of an apnea/hypopnea episode, decreasing to 8.5 ± 12.5% with nCPAP treatment. MB time index per hour of sleep was reduced from 13.5 ± 10.2 min prior to treatment to 4.6 ± 5.5 min under nCPAP (p < 0.05). Conclusions: In OSAS patients, MB episodes often appear at the termination of an apnea/hypopnea episode. In many cases, MB episodes can be markedly reduced by nCPAP treatment. When patients on nCPAP complain of dry mouth, appropriate measurements should be performed to verify MB.

All these papers (and others) document more apneas during mouth breathing - as measured druing a PSG. My conclusion is that the lower number of apnea found in you observations is the result of their not being identified.

O.

Velbor wrote: appear to view a mouth leak as something qualitatively different from a mask/skin leak. I'm not convinced that this is so. I assume that pressure equilibrates very quickly within the system (machine / tubing / mask / airway), and that the machine - and the body - has no way to distinguish between mouth leak and mask/skin leak. The pressure in the mask is the pressure in the airway, and this is so whether or not there is leak, and whether leak is from the mask/skin interface or whether it is from the mouth. If the machine is able to compensate for leak by maintaining pressure in the tubing / mask, then that same pressure is also present it the airway. Leak, from whatever source, either overwhelms the entire system, or it does not.

Given what you believe, above, it may be fruitless to offer this, but I think you are over-simplifying where that is not appropriate. Imagine that you take 2 very large trash bags, and fill each one with air through a small hole in the bottom, having first closed the top securely. Now prick a pinhole or small cut in one of the bags, but completely open the top of the other. In which bag might you be able to maintain the bag's filled/pressurized appearance by continuing to supply air through the hole in the bottom?

The size of the "intake" and "outtake" matter. A mouth is huge compared to the little space of a mask leak. It is reasonable to suppose that you might push enough additional air through a 1" hose to compensate for mask leaks. It is not reasonable to suppose that you could increase the amount of air through a 1" hose to compensate for an open mouth.

But connected parts of the system will be at the same pressure.

Under this theory, you don't need to wear a mask at all. The whole house is connected to the PAP machine when it's running unattached to a mask, so just shut the door and breathe normally.



I mouth tape, but I don't think everyone should. I have no data to offer here, but the reason I mouth tape is very simple. IT WORKS. I feel dramatically better when I tape than when I don't. Do you have any recollection as to whether you felt the same with and without taping?

In any event, this was very interesting, and I appreciate the time you spent in presenting this data. Thank you. As you say, it is a single individual's record, so it might not apply to everyone, but hopefully it will cause people to think a little more critically about what they are doing.

Ozij,

Thank you for your, as always, well researched and comprehensive response. I will not pretend to have thoroughly reveiwed your references. The quotations you selected, tough, seem to say 1) patients with obstructive apnea tend to be mouth breathers, and 2) apneic episodes tend to conclude with an oral gasping breath. Yes. I'm not sure how this is quite relevant to the context of our overall discussion.

LoQ,

Thank you also for your consideration. I would respond that the "size" of the leak opening (but not necessary the location of the opening) is relevant to the extent that it determines the leak rate, which determines whether the machine either can, or cannot, maintain pressure. ResMed specification indicate that their machines can compensate for leaks below 0.40 l/s, that things become "iffy" above that, and that all bets are off above 0.50 l/s. That it precisely why leak "problems" should be quantified. A leak within the machine's ability to compensate may be problematic for any number of reasons, but not in terms of interfering with machine sensor function or maintenance of theraputic airway pressure. A fully open mouth needs to be closed for successful therapy; that's what chin straps are for. Taping, however, is often used (as I used it) to remedy small and transient exhalatory lip openings; these might or might not cause situations which exceed machine capability.

Only if mouth leak rate is within the machine's ability to compensate does the data I presented raise an issue of possible concern. Can taping, under those circumstances, worsen the freqeuncy of apneas and the time in apnea? That is my interpretation of the data. Ozij, on the other hand, posits (if I may be so bold as to summarize your position) that what I interpret as a "worsening" of apnea when using tape is really a reflection of the machine's failure in the presence of mouth leak and without tape to accurately recognize and record (and perhaps to "treat") apneas, even if that mouth leak is within machine specifications.

Regards, Velbor

Velbor wrote: Taping, however, is often used (as I used it) to remedy small and transient exhalatory lip openings; these might or might not cause situations which exceed machine capability.

I hesitate to say this, because as Rooster or someone is always pointing out, without external data, we really don't know what we do in our sleep. But in my case, I'm pretty sure my mouth is falling wide open.

Are you sure you only have "small and transient exhalatory lip openings"? Before I started taping, I was pretty sure I didn't even have lip flutters. I was so wrong.

An interesting thing I noticed as my therapy progressed. Initially, I would wake up several times a night because of the air rushing out of my mouth from the therapy. It was disruptive enough to wake me up every time. After a few weeks, I noticed I wasn't waking up any more. My first assumption was that I had learned to keep my mouth shut. Many people here believe they have done that; I suspect at least half of them just aren't noticing any more that their mouth is open. That was me. I was still opening my mouth, it just wasn't waking me anymore.

Wrong conclusions are so easy to jump to.

Perfect example.
That's why data is so critical.
Particularly data that challenges our assumptions.
That's why methodical process and careful analysis are critical.
Of course, (after Ozij), we must understand the limits of our lab instruments, too.
But science begins with using trusted instruments and trusting our observations.
Then we follow wherever the data takes us.
Velbor

ozij wrote:You are basically saying that the only problems with mouthleaks is the fact they dry your mouth, and cause arousals. According to your understanding they have no impact on the machines' ability to identify apnea, hypopnyea or flow limitations, and therefore have no effect on the machines' abililty to give proper treatment.

Velbor wrote:Nor have I seen documentation that during an apneic event, persons reflexively attempt to open their mouths.

Mouth Breathing Compromises Adherence to Nasal Continuous Positive Airway Pressure Therapy

No distinction was made between mouth breathing and mouth opening at the end of apnea caused by choking.

Baseline Demographics and Sleep Characteristics
We studied 51 patients (4 women). MBs and NBs did not differ with respect to age, sex, body mass index (BMI), neck circumference (NC), ESS, AUDIT score, TST, TST while supine, rapid eye movement (REM) sleep, wakefulness after sleep onset (WASO), sleep efficiency, RDI while supine, or pulse oxygen saturation (Spo2) awake. Meanwhile, total arousal, respiratory arousal, RDI, and an oxygen desaturation index of 4% (ODI4) were higher in MBs, but the amount of delta sleep was lower

.

Effect of nasal or oral breathing route on upper airway resistance during sleep

During wakefulness, upper airway resistance was similar between the oral and nasal breathing routes. However, during sleep (supine, stage two) upper airway resistance was much higher while breathing orally (median 12.4 cmH2O·L–1·s–1, range 4.5–40.2) than nasally (5.2 cmH2O·L–1·s–1, 1.7–10.8 ) . In addition, obstructive (but not central) apnoeas and hypopnoeas were profoundly more frequent when breathing orally (apnoea-hypopnoea index 43±6) than nasally (1.5±0.5).

Obstructive sleep apnoea and oral breathing in patients free of nasal obstruction

In conclusion, apnoeics spent more time breathing orally and oro-nasally than simple snorers, and the apnoea/hypopnoea index is a major determinant of the time spent breathing orally and oro-nasally.

Mouth Breathing in Obstructive Sleep Apnea prior to and during Nasal Continuous Positive Airway Pressure

The lowest SaO2 measured was 82.9 ± 4.7%, and increased to 87.5 ± 2.7% under nCPAP. The mean nCPAP was 7.8 ± 1.6 cm H2O. MB event index per hour of sleep decreased from 35.2 ± 19.7 prior to treatment to 5.0 ± 5.2 under nCPAP (p < 0.01). In 52.2 ± 27.4% of obstructive respiratory events, MB started at the end of an apnea/hypopnea episode, decreasing to 8.5 ± 12.5% with nCPAP treatment. MB time index per hour of sleep was reduced from 13.5 ± 10.2 min prior to treatment to 4.6 ± 5.5 min under nCPAP (p < 0.05). Conclusions: In OSAS patients, MB episodes often appear at the termination of an apnea/hypopnea episode. In many cases, MB episodes can be markedly reduced by nCPAP treatment. When patients on nCPAP complain of dry mouth, appropriate measurements should be performed to verify MB.

All these papers (and others) document more apneas during mouth breathing - as measured druing a PSG. My conclusion is that the lower number of apnea found in you observations is the result of their not being identified.

O.

Looks like I need to do some reading...

A cursory look has me thinking I should try taping my mouth closed and using my FFM and see what my numbers do...

Ozij had commented on 7 March (bolding mine) that:

Ozij wrote:Velbor, I assume that lip flutters, or open mouth breathing, disable the machine's ability to identify an apnea or a hypopnea. As we know, that identification is based on various properties of a series breaths taken previously, (2 minutes on a Resmed) and to my untutored mind, it seems these parameters change dramatically once part of the air blows out of your mouth. Having seen many mouth leak reports in my data (with nary an even occuring during leak time) I agree with you that a mouth leak does not effect therapy by disabling proper pressure levels - we know some machines can, and do automatically compensate up to a certain level. But a mouth leak (unlike a mask leak) does not let the proper pressure build up against the obstruction in your upper airway - the machine supplies the compensating pressure, but the pressure does nothing against the obstruction - the air pours out of your mouth. No events reported deos not mean no events happening.

Ozij is known to be a careful and thoughtful contributor. A fuller examination and analysis of my data turned out to be possible, and will hopefully be useful. First, here is an example (from an AutoScan “export” of detail data to Excel; Activa mask, no mouth tape, fixed pressure 11.0) of an apnea recognized and reported during a very high leak “lip flutter” cycle (outside the machine’s stated specifications):



I have already presented data which quantifies the percentage of time spent at various levels of leak. In theory, ASSUMING that the occurrences of reported apneas are INDEPENDENT of current leak rate, the distribution of apneas at various leak rates should approximate the distribution of time spent at each level of leak. I think that Ozij would NOT agree with that assumption: she would presumably posit that there would be fewer apneas recognized and reported, as mouth leak increases. I also DISAGREE with this assumption, but in the opposite direction: I would posit that increased levels of “lip flutter” leak might occur during deeper sleep, when an increased rate of apneas might also be expected, resulting in relatively more apneas reported as leak increases.

Here I present additional, updated leak distribution information (20 nights using the Activa mask with NO mouth taping at a fixed 11.0 pressure). Further, I have reviewed my data to similiarly quantify what the leak rate was when each reported apnea occurred:



Unfortunately for the analysis (though fortunately for me), the number of apneas is relatively small (the average nightly AI was 0.2), making it difficult to draw statistically definitive conclusions. While recognizing this limitation, it nonetheless appears that there are disproportionately more apneas recognized and reported during times of leak, and disproportionately fewer apneas during times of zero leak (which is the median condition).

In summary, while not constituting definitive proof, I interpret this data as supporting the ability of the ResMed S8 machine to sense apneas during leak conditions (including mouth leak) within (and to some degree, beyond) the machine’s specifications.

Velbor

In my original posting, I presented data and proposed three counter-intuitive conclusions. Most discussion so far has been about the conclusion that mouth taping can result in decreased efficacy.

I would, however, now return to another of the conclusions from my original data: that xPAP interfaces (or at least the ResMed Activa and UMFF masks) are not merely passive devices to provide a comfortable and reliable facial seal, but in fact function as active components, influencing the results of xPAP therapy.

My original data demonstrated that in AutoPAP mode (6 to 20 cm), the 20-night average 95th centile pressure results differed by almost a full cm. between the masks:

Activa: 11.0 (standard deviation 0.2)
UMFF: 11.9 (standard deviation 0.7)

The statistical strength of this difference is inconclusive. Other differences between the two masks, in terms of pressure distribution over time, and distribution of pressures at which apneas occurred, were also presented.

At constant CPAP pressure of 11.0 cm, results for ten nights with each mask have already been presented. Updated data presenting averages for twenty nights using each of the two masks is now available.

(Previous data with the Activa and mouth taping is repeated for completeness. I have not, and do not plan to, obtain further data with mouth taping. If this were a clinical trial, I think it would be unethical, based on interim findings, to continue taping research subjects. It seems proper to treat myself in the same way. Thus, the additional comparative data provides little new information regarding the possibly deleterious effect of mouth taping when using the Activa mask.)



My choice of 11.0 cm was based on the 95th centile APAP value for the Activa without taping. As might be expected from the APAP suggestion that the UMFF needed to be operated at a higher pressure, the therapeutic results of the UMFF were inferior to those of the Activa without tape. Once again, the statistical strength of this difference is not sufficient for proof. Still, the four-fold difference for AI, and the greater than three-fold difference for time in apnea, is notable.

More striking, perhaps, is a scatter-plot of the mask performance differences



While the statistician would hope for “tighter” and non-overlapping clusters, the visual impact of the superiority of the Activa without tape over the UMFF (and even more so over the Activa with mouth taping) is impressive.

The updated “leak performance” of the masks (using minute-by-minute leak data to categorize the percent of time in each leak category across all nights) demonstrates that the Activa without tape is operating within machine specifications more than 95% of the time. Comparison of the Activa with and without tape documents that face seal is almost perfect, and that Activa leak without tape can be assumed to be almost entirely “lip flutter” from the mouth.



Whether or not the extent of Activa leak without tape is or is not acceptable is debatable. If one believes that the machine is able to function as claimed (maintenance of therapeutic pressure and accuracy of sensors) for leak rates below 0.40 l/s, then the overall time of POTENTIAL substandard function (in which direction?) remains less than 5%, which is arguably tolerable, particularly in view of the efficacy achieved. However, four individual nights with 95th centile leak of 0.40 l/s or over were observed; the acceptability of this situation is questionable. As my data demonstrates, mouth taping is NOT a viable option for me for managing the leaks. But the point of my posting this data is not to evaluate my own situation – it is to emphasize the NEED for data to evaluate xPAP therapy and to make knowledgeable decisions.


To summarize, the conclusion I would draw from this data is that it supports the suggestion I presented at the outset: that the Activa and UMFF masks are not simply passive components, but actively (though the mechanism is unclear) affect CPAP performance. Presumably, this is true not only for me, and may well be true not only for these specific masks.

If this is accepted, then a corollary would be that the concept of a single physiologic “titration pressure,” independent of interface equipment, and derivable from a single-night PSG, is overly simplistic.

The practical corollary which follows would be that masks (or mask types) should not be switched indiscriminately; that DATA is needed to confirm efficacy and perhaps to adjust pressure.

Regards, Velbor

My latest “experiment” (3/09 – 4/09) has been a return from fixed pressure back to AutoPAP (ResMed S8 Vantage) and, based on the successful 11.0 cm pressure with the Activa, and with the intent of lowering the median pressure to reduce leak (though from 11.0 down to only an average of 10.7), I decided to utilize a minimum pressure of 10.0 cm. Maximum pressure was left at 20 cm, since my previous experience at that setting demonstrated no problems. Results (with previous values at 6-20 cm repeated for comparison) were as follows:



Before proceeding to an analysis of the new data, it should be pointed out that, once again, my data provides no evidence to support the "conventional wisdom" that it is preferable to set the minimum pressure just under the "titration pressure". The difference in respiratory disturbance indices when the machine is set "wide open" (6-20) or when the minimum is raised (10-20), regardless of mask, is of no clinical relevance nor of any statistical significance. In fact, the Activa efficacy parameters are slightly worse, and the UMFF parameters slightly better, than they previously were, though the differences are virtually meaningless.

Focusing on the new data at settings of 10-20 cmH2O, the Activa (without mouth taping) once again provides superior therapeutic efficacy "numbers" than does the UMFF mask, though this time the differences between the two masks are so small as to be both clinically and statistically irrelevant. Indeed, based on this data, one might accent the similarities as legitimately as the differences. This fact is particularly cogent when considering my much higher unintentional leak rate on the Activa (due to oral exhalation), as will be discussed later.

Still, it is interesting to view the efficacy parameters in a scatter plot:



Eighteen (90%) of the 20 Activa without tape nights are "within the box" (arbitrarily, AHI ≤ 10, AI ≤ 1), while only eleven (55%) of the UMFF nights are similarly situated. (Not all individual data points are visible, due to overlapping.)

While the efficacy differences between the two masks are small in this dataset, pressure-response differences between the two masks remain notable and significant. While the median pressures are almost identical (Activa = 10.7, UMFF = 11.0), the 95th centile pressures (as also noted previously at 6-20cm), are over a cm apart (Activa = 11.6, UMFF = 12.8. The difference of 1.2cm exceeds the 1.0 sum of the standard deviations.) Differences in maximum pressures are even more striking, with the Activa never exceeding 12.6 cmH2O, while the UMFF reached a maximum pressure of 16.8 cmH2O, 4.2cm higher.



While both masks display a modal pressure in the 10.0 – 10.9 cm range, the pressure profile of the Activa is narrow, at or above 12.0 only 1.7% of the time. The UMFF, however, spent notably more time at higher pressures, at or above 12.0 cm 17.1% of the time.



Similarly, apneas are seen with the Activa principally while pressure is in the 10 and 11cm ranges; only 4.7% of apneas (n = 3) were in the 12cm range, with none higher. Using the UMFF, only about half (54.6%) of the apneas occur below 12cm pressure, while the remaining apneas gradually taper off, with a single apnea appearing at the maximum pressure of 16.0 cm

Looking individually at each mask’s performance in terms of pressure distribution over time, and pressure-at-apnea distribution, we see:



For the Activa, the two plots are virtually identical.



For the UMFF, the time-pressure profile drops off more quickly than does the pressure-at-apnea profile. The significance of this observation, if any, is not clear. Still, the correspondence of curve shapes should give some pause for thought to detractors of the performance of the ResMed algorithm at pressures above 10cm.

Leak characteristics of the two masks were as follows:



These statistics are not markedly different from those previously presented. The UMFF once again displays an excellent leak profile. Due to my “lip flutter” exhalations, the Activa’s leak profile is far less desirable. Still, machine specifications are “exceeded” less than 2% of the time. In addition, none of the concerns I had previously expressed (at 11cm constant pressure) have reappeared: there have been NO individual nights with a 95th centile leak at or above 0.40 l/s (the highest has been 0.38, twice), and there have been only three nights with leak at or above 0.50 l/s (for 0.1% of total therapy time; maximum 6 minutes during any night).

A detailed examination of Activa leak status during apnea episodes, as presented previously, does not in this dataset reveal any clear pattern regarding any relationship between leaks and apnea occurrences (see subsequent note to ozij). During the (median situation) 57% of the time with NO leak reported, 64% of apneas occurred. Only a single apnea (1.6% of the total) occurred during the 1.8% of the time with “significant” leak. The number of apneas is too small to provide statistical significance.

SUMMARY In summary, the thrust of the data presented in this experiment, reinforcing the observations made earlier, is that there are significant performance differences between the Activa and the UMFF masks when used under identical conditions. While efficacy differences (i.e., AHI, AI, HI and time-in apnea) are smaller than previously reported, pressure-profile differences remain notable, though of uncertain relevance. While the presentation of my OSA may possibly include some unique characteristics, it seems likely that these differences may affect other users as well. If so, mask selection cannot be regarded as simply a "comfort" issue.

I have no information as to whether these properties are unique to these particular masks, or whether they may reflect generic performance differences between nasal and full-face masks. Still, the concept of a single physiologic "titration pressure", determinable by a single-night PSG and without reference to the interface equipment, again appears to be overly simplistic.

Finally, I expect that this will be my last posting of new data in this thread, though I anticipate that I will soon post here a "meta-analysis" overview of all the data. The choice of mask appears to be more predictive of results than the choice of machine settings. I am also beginning a project of similar data collection and analysis using the Respironics M Series AutoPAP. While we must be VERY careful "comparing" machines from different manufacturers – often the same terms are used with different meanings – preliminary indications suggest that the data will be quite interesting! I will plan to begin a new thread with that information.

For ozij, a special note We share a respect for data, a respect for statistical analysis, and a respect for each others’ honesty and integrity. I therefore must present some details, at least for you.

I again did the leak / pressure correlation analysis for the Activa mask without mouth taping, which last time supported my hypothesis. This time, the data does not support the position I had hypothesized.



There is an excess of apneas reported during low-leak conditions, with respect to the time spent in each leak category. This arguably supports your hypothesis of decreased apnea reporting during periods of higher leak.

Still, there are the same weaknesses as last time, when the data favored my view. While there are now more apnea occurrences, the data remains relatively sparse, and even small variations (and I typically engender major variations!) would skew the outcome markedly.

I remain convinced that there is no a priori reason to doubt the accuracy of apnea sensing and reporting while leak is within the machine’s specifications, whether the leak is at the mask-skin interface or whether the leak is from the mouth. As I proposed in another thread in another context, when leak is within machine specifications, pressure throughout the system, from the blower exit through the tubing and mask and within the airway is, for all practical purposes, instantaneously equilibrated, regardless of the location of the leak.

I remain unable to explain the previously presented efficacy differences between Activa with mouth taping and Activa without mouth taping. I am similarly at a loss to explain the differences I have just presented between the UMFF mask and Activa (although, as I have pointed out, efficacy similarities may be more impressive than differences!). I respect your hypothesis as a thoughtful, innovative attempt at explanation. Although I can offer no alternative explanation, I am not convinced by yours.

But I am impressed by your mind, and your suggestion cannot be idly dismissed!!

With apologies for the length of this post (obviously, I find it fascinating!), I hope that this material is of some interest, and perhaps even of some use. Blessings, Velbor

FINALE : META-ANALYSIS

As I noted in my previous post, “The choice of mask appears to be more predictive of results than the choice of (ResMed S8 Vantage) machine settings.” The data from the individual experiments, already presented in this thread, upon which I base that assessment, follows:



Thus I have compiled data from experience with five machine settings and three mask variations:



Combining the original nightly data for all of these individual experiments by mask type, without differentiation as to machine setting, provides the following efficacy results:



All efficacy numbers consistently improve as we move down the chart, though there are varying levels of significance, and there is certainly overlapping of value ranges.

Comparing the Activa used WITH mouth taping to the Activa used WITHOUT mouth taping, the difference between the average values for all parameters except HI exceed the sum of the standard deviations. In addition, for AI and % time in apnea, the average values for the Activa WITH tape are well above the maximum range for the Activa WITHOUT tape. These differences may be considered significant.

Comparing the UMFF to the Activa used WITH mouth taping, the difference between average values of AI and % time in apnea exceeds the sum of the standard deviations. In addition, for AI only, the average value for the Activa WITH tape is above the maximum range for the UMFF. These differences may be considered significant.

Comparing the UMFF to the Activa used WITHOUT mouth taping, there are no parameters for which the difference between average values exceeds the sum of the standard deviations, and in no case is any UMFF average value above the range of values for the Activa WITHOUT tape. These differences, therefore, cannot be considered statistically significant, or alternatively stated, my efficacy results with these two masks are statistically equivalent.

I provide the customary scatter plot of all of the data points:



and trust that you will arrive at your own conclusions.

Finally, in what I expect to be my concluding post for this thread, I take the liberty in the context of this meta-analysis to restate the three counter-intuitive observations with which I began this thread:

 Narrow AutoPAP ranges are not necessarily better than wider ranges.
 Mouth taping can adversely affect therapeutic performance; while it reduces or eliminates leak, it can also degrade therapeutic effects.
 Masks are not just comfort interfaces; different masks differently impact CPAP performance, in terms of pressure profile, and in terms of therapy results.

Blessings, Velbor

Effect of mouth leak on effectiveness of nasal bilevel ventilatoryassistance and sleep architecture
Eur Respir J 1999; 14: 1251±1257
H. Teschler*, J. Stampa*, R. Ragette*, N. Konietzko*, M. Berthon-Jones**

ABSTRACT: Mouth leak is common during nasal ventilatory assistance, but its
effects on ventilatory support and on sleep architecture are unknown. The acute effect
of sealing the mouth on sleep architecture and transcutaneous carbon dioxide tension
(Ptc,CO2) was tested in 9 patients (7 hypercapnic) on longterm nasal bilevel ventilation
with symptomatic mouth leak.
Patients slept with nasal bilevel ventilation at their usual settings on two nights in
random order. On one night, the mouth was taped closed. Leak was measured with a
pneumotachograph.
Median leak fell from 0.35‹0.07 (mean‹SEM) L.s-1 untaped to 0.06‹0.03 L.s-1 taped.
Ptc,CO2 fell in 8/9, including all hypercapnic patients. Across all patients, the mean
Ptc,CO2 fell by 1.02‹0.28 kPa (7.7‹2.1 mmHg) with taping (p=0.007). Arousal index fell
in every patient. Mean arousal index fell from 35.0‹3.0 to 13.9‹1.2.h-1 (p<0.0001), and
rapid eye movement (REM) sleep increased from 12.9‹1.5% to 21.1‹1.8% sleep time
(p=0.0016). Slow wave sleep changed inconsistently, from a mean of 13.1‹1.6% to
19.5‹2.2% of sleep (p=0.09). Sleep latency and efficiency were unchanged. In four
healthy volunteers ventilator-induced awake hypopharyngeal pressure swing during
timed bilevel ventilation fell by 35‹5%.L-1.s-1 of voluntary mouth leak (p<0.0001).
Mouth leak reduces effective nasal bilevel ventilatory support, increases transcutaneous
carbon dioxide tension, and disrupts sleep architecture.

My emphasis....
From the above:

In the only study to date to directly measure the effect of
mouth leaks on effectiveness of ventilatory support, CARREY
et al. [6] investigated the effect of voluntary mouth
opening on diaphragm electromyogram (EMG) activity
during awake nasal intermittent positive pressure ventilation
in 5 subjects. With the mouth closed, diaphragm
EMG dropped to 15% of unassisted control, but with the
mouth open, returned to 98% of unassisted control, implying
complete loss of ventilatory support

There is an urgent clinical need for a safe and effective
method for preventing mouth leaks during bilevel ventilatory
support. Taping was used in this study in preference to
other ways of sealing the mouth such as a full face mask,
because existing full face masks have higher deadspace,
leak, and discomfort than nosemasks, and arousal data
would be uninterpretable. However, the authors do not at
this stage advocate taping the mouth for indiscriminate
long-term home use, because of the risk of asphyxia in the
presence of nasal obstruction, machine or power failure, or
regurgitation.
Since performing the present study, ~30 carefully chosen
patients, relying on nocturnal support for ventilation but
unable to be satisfactorily treated due to mouth leak, unable
to tolerate a full face mask, highly alert, with no history
of regurgitation, and faced with the alternatives of
tracheotomy or untreated respiratory insufficiency, have
elected to use tape at home for periods of up to 6 months.
The tape is well tolerated and no adverse events have been
observed. In one patient, taping the mouth caused the
Ptc,CO2 to fall acutely from 8.91 kPa to 5.32 kPa (67
mmHg to 40 mmHg), and it was necessary to reduce the
ventilator settings to compensate.

All emphasis mine.
By the way, the study by OKayed by an ethics commitee.

Interesting mouth taping benefits for asthmatics:
http://www.sciencedirect.com/science?_o ... 2034b661ef

Very nice thread, Velbor! Thank you!

-SWS wrote: Very nice thread, Velbor! Thank you!

Thank you, -SWS. Your compliment means a lot to me.

ozij wrote:Effect of mouth leak on effectiveness of nasal bilevel ventilatoryassistance and sleep architecture

Ozij, You are a veritable fountain of important literature references! Thank you. I would make a few comments regarding the article you cited.

It is fascinating that such a “taping” study would be done! No argument from me that mouth leak is undesirable, that it should ideally be minimized, and that mouth leak CAN interfere with therapy and with sleep itself.

The patients studied in the citation were unfortunately not being treated for Obstructive Sleep Apnea, so we have no information about the effect of taping on the index parameters we commonly focus upon. Just how “atypical” my own numbers are, with mouth taping associated with worsened efficacy, remains unknown.

The article provides the following leak statistics table:



This suggests that they are reporting net leak excluding normal mask venting, since the taped values (0.03 – 0.09 L/s) are far too low to include mask venting (Activa venting @ 10cm, e.g., = 31 L/m = 0.52 L/s). The untaped control values are extremely high numbers, significantly exceeding the standard ResMed machine tolerable leak specification of 0.40 L/s for large amounts of time. As I have indicated, my own Activa untaped median leak was typically zero, and 95th centile leak rate was typically well below 0.40 L/s.

I was most interested in reading that the average arousal index fell (though sleep efficiency was unaffected). While EEG arousal measurements are beyond what I can do at home, a (poor) surrogate which is easily measurable is the pattern of arisings from bed.

In addition to the formal data I have presented previously, I also recorded (to the nearest 15 minutes) the time I started xPAP each night, and the time of each arising from bed. (Prostate issues complicate my sleep patterns.) Composite averages (and ranges) for each mask configuration follow (with no apologies for not providing standard deviations on this inherently “soft” data):



Edit: Minimum UMFF Bed Hours should read 6.75.

I am not prepared to make any evaluation of the difference in bed hours for each mask. I suspect they are not significant. Even if they are, whether they reflect the quality of sleep provided by each mask, or the NEED for sleep while using each mask, is unknown. The longest bed time is with the Activa with tape (lowest leak); I have no idea what that means.

The smallest average number of arisings from bed again occurs with the Activa and taping (lowest leak). Related is the calculated “Up Index”, the number of arisings per hour (Times Up / Bed Hours). Here also, the Activa with mouth taping has the smallest or best value, though again I suspect there is no significance to the differences. Also related is the calculated “Average Hours Sleep Segment” length (Bed Hours / Times Up +1).

Average values for the longest block of uninterrupted sleep each night are similarly fairly close to each other, though again, Activa with taping (lowest leak) is the best. Related is the calculated “Maximum Continuous Sleep as a Percent of Total Bed Hours”, for which the values are very close and unlikely to be significantly different.

In summary, I conclude from my data that if my mouth leak with the Activa without taping has any deleterious effect, it is not expressed in terms of these quantifiable “sleep quality” parameters. (And similarly, the documented deleterious effect of mouth taping on my respiratory event indices is not reflected in these “sleep quality” parameters.)

Regards, Velbor

The index parameters you are focusing on, Velbor, are not the only varialbles we should consider.
We cannot assume the efficacy indices are direct indicators of quality of sleep.

This study may sugges a different expalantion of your own taping data:

Taping leads to more REM and less arousals. For the majority of us, more apneas occur in REM - assuming your obstructive sleep apnea is just that, we can say that: Taped, you dream more and have more apneas during that time. Untaped you have more arousals, some of which actually make you arise...


I linked to this study also because of the way it discusses various possible explanations of how mouth leaks cause the drop in ventilation - the physics of the lowered hypopharyngeal pressure is beyond me - but not the fact.

In conclusion, during bilevel ventilatory support, severe mouth leak can cause an important reduction in the ventilator-induced pressure swings at the hypopharynx, particularly if the nose is congested. High mouth leak is very distressing to wake control subjects and is a likely cause of severe sleep disruption. On the first night, taping the mouth closed led to more than halving of the arousal index and a worthwhile increase in rapid eye movement sleep, with no increase in sleep latency or reduction in sleep efficiency, and led to a reduction in transcutaneous carbon dioxide tension in hypercapnic patients. There is a need for routine assessment of the degree of mouth leak in nasally ventilated patients, with a view to preventing mouth leak where possible.