Introduction to Leak
Air is sucked into the CPAP machine from your room. It is drawn in through a filter, and is blown out at a controlled flow rate to provide the desired pressure. Air blown out of the CPAP machine eventually leaves the machine - tubing - interface - person system and returns to your room. There are four possible routes by which air is returned to the room:
(1) Intentional venting. Virtually all commercial CPAP interface systems are designed so that air ALWAYS blows through them. This enables the moisture-laden, oxygen depleted, and carbon dioxide-rich air which you exhale into the interface to be continually "washed out" so that it is not re-inhaled. Your exhalations become part of the venting airflow. This venting is part of the interface design, with greater volumes of air being vented at higher pressures. Most manufacturers provide information as to how much leak will occur at various pressure levels. While often referred to as "leak," this normal, intentional venting should not be confused with "unintentional" leak. Intentional venting is the only DESIRABLE way for air to return to your room.
(2) Seal leak. Virtually every interface device involves some sort of contact, a "seal," between a plastic "cushion" or "pillow" and the user's skin. Air can easily leak out through these seal contact points, particularly as the user moves around during the course of the night, or as pressure increases (as with an AutoPAP). This is the first type of "unintentional leak," which should ideally be minimized.
(3) Mouth leak. This is the second type of "unintentional leak," which also should ideally be minimized. Since the mouth and the nose are connected in the pharynx, unless a "Full Face Mask" (or other special interfaces which cover both the mouth and the nose) is used, air which enters the nose may leak out of the mouth. This can occur in several different ways (or in combinations) which may more-or-less affect the effectiveness of CPAP, and which may require different strategies for dealing with the leak:
(a) "mouth breathing," in which unpressurized room air is inhaled - and exhaled - through the mouth,
(b) "open mouth flow," in which air pressurized air entering through the nose more-or-less continually escapes through an open mouth, or
(c) "mouth exhalation," in which pressurized CPAP air which is inhaled through the nose is totally or partially exhaled through the mouth.
(4) A defect in the equipment, or equipment defectively assembled. This is always bad, and should be guarded against by frequent inspection, and periodic listening and feeling for extraneous airflow.
What Harm Do Leaks Cause?
There are three mechanisms by which unintentional leak can detract from CPAP therapy. One or more of these mechanisms may be present. None are "all or nothing" phenomena; the nature, extent and freqeuncy of leak will determine the negative effect - if any - of the leak. It may accurately be stated that, "Leak is a problem only if leak is a problem."
(1) Disruption of Sleep. Leak involves the blowing of air where air "should not" be blowing. This is usually related to seal leak, but may also be applicable to mouth leak. If air is blowing across the face or into the eyes, this is likely to arouse or awaken you from sleep. Blowing air also creates sound, ranging from rushing torrents to "musical" tones when the edges of the plastic cushion vibrate. These sounds too can arouse or awaken you, interfering with sleep.
(2) Drying of the Airway. Mouth leak has the special disadvantage of the leak taking place AFTER the pressurized air is already within the body. While all leak involves additional airflow, mouth leak produces additional airflow blowing WITHIN the nose and mouth. This often results in drying of the membranes. If the nasal membranes become dried and irritated, increased resistance to breathing may ensue, with resultant congestion. Drying of the mouth may be uncomfortable, and add to sleep disruption.
(3) Pressure Loss. This problem is sometimes overemphasized and often misunderstood. To understand it properly, a bit of physics needs to be reviewed. In a small, passive, closed, pressurized system (think of a car tire, perhaps), the pressure is always everywhere the same. If a "small" leak develops in the system, the pressure drops, though it still remains the same everywhere in the system, regardless of the site of the leak. CPAP is a "small" system, and for all practical purposes the pressure everywhere - in the tubing, in the mask, in your airway - is always instantaneously the same, regardless of where the leak originates. But CPAP is not a "passive" system. The technology works by increasing or decreasing air flow in order to maintain the target pressure the machine is trying to deliver. This is important: CPAP "uses" airflow but CPAP "works" by maintaining airway pressure. Leak is the loss of airflow, but leak does not necessarily result in a loss of pressure. Modern CPAP machines are robust, and can compensate for significant loss of airflow from leaks by increasing their output airflow. In this way they maintain their desired pressure, even in the presence of substantial leak, and regardless of the location of that leak. (Thus a "mouth leak" is not necessarily any more relevant in terms of pressure than is a "seal leak.") However, if the leak is too large, it may overwhelm the machine's ability to compensate and maintain pressure. This is arguably the most severe and important consequence of leak, though its frequency may be exaggerated. It takes a LOT of leak to result in failure of the machine's ability to maintain pressure and/or the ability of the machine's sensors to accurately detect pressure and flow information. When this happens, however, the effectiveness of CPAP therapy can be markedly compromised.
Respironics documentation summarizes the above information nicely: "System leak is a combination of intentional and unintentional air leak. Intentional leak is the expected leak at the exhalation port. Some leak is required to minimize CO2 rebreathing. Unintentional leak occurs around the patient interface. If there is a large increase in the amount of leak .... the patient may need a mask refitting. Leaks that should be fixed include leaks into the eyes, leaks that bother the patient, or leaks that affect pressure stability."
How Do I Know If I'm Leaking?
Unfortunately, we're not always aware of whether or not, or of how significantly, we may be experiencing unintentional leakage. The blowing of air, or the sound created by blowing air, doesn't always awaken us. We may not be aware of, or on awakening not remember being aware of, a dry mouth caused by mouth leak (particularly if the leak is intermittent).
Fortunately, chances are that if leak is not frequent enough or large enough to cause symptoms, it's probably not significant enough to be of concern. Still, the possibility of an unrecognized large leak, which might interfere with effective therapy, cannot be discounted entirely.
Many CPAP machines provide an audible "Leak Alert" alarm function. If given a choice, and unless there is a particular reason NOT to use it, this option should generally be turned "on."
Many ResMed (S7, S8) machines offer a "Leak Alert" which can be enabled or disabled. "When enabled, leaks > 0.7 L/s for > 20s result in an audible alert and a high leak message in the LCD. Note: when Leak Alert is enabled, SmartStart/Stop is automatically disabled." We will talk more, in a subsequent section, about what these numbers mean. For the present, we might simply note that 0.7 liters per second (equal to 42 liters per minute) is a sizeable airflow. Given that a typical adult breath moves about 500 cc of air (Tidal Volume), and that we might perhaps be taking 10 breaths per minute, the amount of air moved by normal breathing is about 5 liters per minute, or 0.08 liters per second. Thus, the machinery quite reasonably becomes concerned when leakage accounts for over eight times more air movement than is produced by our breathing. Particularly for AutoPAP machines, ResMed notes that "When the leak exceeds this level, the autotitrating algorithm ceases to perform optimally."
Again, simply to keep this number in context (pending more detailed discussion below), ResMed notes that "Generally, a leak rate of more than 0.4 L/s (24 L/min) is associated with patient discomfort, disturbed sleep, and reduced efficacy of treatment."
Many Respironics machines (M-Series) also provide a "Mask Leak Alert" feature: "Mask Leak Alert – This flashing text displays on the Active Display screen if the Mask Alert setting is enabled and the device detects an excessive mask leak. .... If this feature is enabled, then the [ ] symbol flashes on the Active display screen if a significant mask leak is detected, and an audible alert sounds." Unfortunately, available Respironics documentation does not define their criteria for "excessive" or "significant" leak.
What Does My Machine Tell Me About My Leak?
More information than simply a “high leak alarm” may be available to those with “data-capable” CPAP machines (which have been configured to display such information). Such machines typically reduce to a single number, and display this number on the machine’s LCD panel, a measure of the severity and frequency of leak over a period of machine use. Discussion is complicated by the employment of different reporting paradigms by the major CPAP machine manufacturers.
Units: Respironics machines generally display leak in liters per minute (lpm or L/m), while ResMed machines generally display leak in liters per second (lps or L/s). This is not a major conceptual problem; the numbers simply differ by a factor of 60.
Centiles vs. Averages: ResMed machines generally display leak as the 95th centile value. These, technically, are the HIGHEST leak values which the machine encountered during the BEST 95% of the night, “best” meaning the portion of the night with the LOWEST leaks. Put another way, ResMed divides the time during which the machine is blowing into 60-second segments, computes the median leak during each 60-seconds, ranks them from lowest to highest, and reports the highest value encountered for 95% of the night.
Admittedly, not very clear. ResMed's practical goal is to encompass, in a single number on the LCD screen, the seemingly impossible task of reflecting some sense of what leak has been like for an entire night. Is the mean or median leak helpful (software generally does display these)? These measures of “central tendency” don’t reflect how severe the leak might have been, for how long. Is the maximum leak helpful (software generally does display this)? A high maximum may have lasted for only a brief period, with low leak during most of the night. What the centile value attempts to reflect (albeit imperfectly) is HOW GOOD or HOW BAD leak was for HOW LONG.
So: with a ResMed machine, if, for example, for the past seven days, your centile leak was 0.10 l/s (or 6 l/m), you know that, overall, for 95% of therapy time, your leak was NO HIGHER than these levels. Not at all shabby. And if, for the past seven nights, your centile leak was 0.40 l/s (or 24 l/m), you know that, overall, for 95% of therapy time, your leak was NO HIGHER than these levels. Well, according at least to ResMed, that’s beginning to enter into questionable territory for some significant portions of the nights. Higher values are plausibly warning signs. An imperfect system, to be sure, but not as totally irrational and unusable as might appear at first glance. (And, regretably, this discussion is a bit of an oversimplification.)
Respironics machines, on the other hand (while reporting a very similar 90th centile leak on their software) presents on its LCD screen AVERAGE leak information (for the latest 7 and 30 days). The limitations of an "average" must be remembered: an average leak of 50 lpm may have been a leak of 40 lpm for half of each night and 60 lpm for the other half of each night - not terrible. Alternatively, that same average 50 lpm may mean a leak of 40 lpm for 3/4 of each night, and an unacceptable leak of 80 lpm for 1/4 of each night: for 2 hours out of 8, leak was excessive. Software is needed to tell which of the scenarios actually occurred.
Granularity: Respironics divides time during which the machine is blowing into 30-second segments, and reports its 30-second average leak data to a precision of 7 lpm (or about 0.12 lps). ResMed reports its 60-second median leak data to a precision of 0.02 lps (or about 1.2 lpm). These differences in level of detail are reflected in the leak number displayed on the LCD panels.
Duration: Respironics machines display average leak values on its LCD screen as running averages for the past 7 days or the past 30 days. ResMed machines display centile leak values on its LCD screen as medians for these same time periods (as well as for six months and one year), but also for the SINGLE most recent night. Given normal night-to-night biological variability, there is some debate as to whether information for a single night is useful (though if one tracks and records the data for each night, it is hard to deny the value of such cumulative information), or whether only when looking at data compiled over multiple nights can valid information be evaluated.
Gross vs. Net: Arguably the most significant difference between the two major manufacturers has been left intentionally for last. Respironics displays as “leak” what is best conceptualized as the TOTAL AIRFLOW leaving the machine. (Remember from the introduction that all air leaving the machine eventually returns to the room, either through the interface ports as intentional “venting,” or as unintentional “leak.”) ResMed, on the other hand, asks the user to identify to the machine WHAT MASK is being used, and the processor SUBTRACTS the expected venting (at the appropriate pressure), displaying as “leak” only the UNINTENTIONAL leak. (Naturally, only ResMed masks are included, so users of non-ResMed interfaces need to select the ResMed mask which is closest in pressure/flow characteristics to the mask they are using.) This distinction makes a great difference in the interpretation of leak numbers presented by the LCD panels of the two machines.
The bottom line: of this long discussion (which will carry over into the topic to follow) is that data-capable machines do make available quantified information about leak. However, how best to utilize that information, and even whether that information is valuable, interpretable and actionable, is subject to much debate. In general, any information is better than no information, but that is true only if the information is capable of being understood and evaluated in a knowledgeable and valid manner.
What Does Software Tell Me About My Leak?
While the LCD screen of a data-capable machine provides one “leak number” (the average or 95th centile, for each of several available time periods), with proprietary software a great deal of additional information about leak becomes available. While Respironics and ResMed software have some major differences, the bulk of the truly significant differences revolve around the different leak “paradigms” reviewed in the previous section. Software from either manufacturer (Encore from Respironics and ResScan from ResMed) offers the following advantages over their respective LCD displays:
(1) Time range selections. The Respironics LCD screen offers average leak information for only the past 7 days and the past 30 days, and the ResMed LCD screen offers 95th centile leak information only for the past night, the past week, month, six months and year. With either manufacturer’s software, information about any group of nights between freely-selectable start and stop dates can be reviewed. This is particularly advantageous for the Respironics user, since leak information for the past night only, which is NOT displayed on the LCD screen, is available.
(2) Additional statistics. Respironics Encore software offers (on the Sleep Therapy Long Term Trend Report), in addition to the average 90th centile leak information, the average MAXIMUM leak, the AVERAGE leak, and the average LARGE LEAK, for the selected date interval. (The time interval can be specified as a single night – the same start and stop dates – so that these statistics for a single night can be viewed.) On the single-night Daily Details report, the number of minutes, and the percent of the night spent, in LARGE LEAK, are displayed, along with AVERAGE leak. On the Therapy Data Summary, the “Average Time in Large Leak per Day” is displayed for the interval. (Each report appears to use different “rounding” rules, so the numbers are not always the same.) These parameters require some explanation. MAXIMUM leak is the highest “instantaneous” leak value encountered – it is NOT the highest 30-second time window value – so this number will not correspond with graphic displays of leak. (In fact, it must be said, this maximum leak number is quite useless.) The AVERAGE leak is clearest: it is the average of the 30-second time window leak values. LARGE LEAK is not well defined in Respironics documentation: it may be what corresponds to the trigger level for their “Mask Leak Alert” feature, or it may be a somewhat lower value. In either case, it reflects what Respironics considers to be an excessively high leak; its occurrence is displayed graphically with thick, dark “warning” lines.
ResMed ResScan software offers, in addition to the median 95th centile leak information, the MAXIMUM leak, and the MEDIAN leak, for the selected date interval. As with Respironics software, information for a single night can be viewed. Unlike the Respironics software, the maximum leak for ResMed is the maximum 60-second time window value, so it does correspond with graphically-presented information.
(3) Summary graphics. Both software packages provide graphic displays enabling an overview of the night-to-night summary leak statistics. These are set up in quite different formats, and while there may be individual preferences for one or the other, they provide by-and-large the same sorts of information. The value of these graphics is in the ability to see, visually, whether leak patterns over successive nights are increasing, decreasing, or, which is frankly most common, quite erratically variable. Still, such graphs can be useful in determining whether, over time, techniques for reducing or mitigating leak (such as those reviewed in the next section) are having the desired effect.
(4) Individual night graphics. Wisely or not, to an extent far exceeding the attention given to the offerings of CPAP monitoring software just listed, the leak graphs for individual nights are by far the major attraction for most users. And with good reason. Each 30-second or 60-second time window recording of average or median leak is displayed, for the entire night. Leak changes can be correlated with pressure changes, and (if there indeed is any correlation, which is debatable) with the occurrence of respiratory events. Brief, isolated leak event spikes can be distinguished from systematic recurrences, and generally ignored. Patterns of leak, which may be amenable to remediation, can sometimes be distinguished. Periodic “hills” of increased and decreased leak may be associated with mouth exhalation which may be more frequent during cycles of deeper sleep or dreaming. But most often, it’s simply interesting, but unactionable.
It must also be remembered, as reviewed in the preceding section, that Respironics presents “leak” as total flow – normal venting plus unintentional leak – while ResMed presents an “adjusted” leak, unintentional leak only, having already subtracted out the expected normal venting for the mask which has been selected (from among its list of ResMed masks). What each of the manufacturers nightly graphs show, therefore, is in fact somewhat different, and must be carefully interpreted.
Since Respironics Encore nightly graphs display total flow, they have a built-in “baseline” of venting. And with an AutoPAP, that baseline changes as pressure varies. This makes the interpretation of Respironics nightly graphs more complex; the interpreter must know and “mentally subtract out” the venting rate of the mask in use to view the unintentional leak. Or more often, the user simply looks at the leak line and subjectively decides (hopefully) that the variability isn’t TOO wide, suggesting that the leak is relatively stable through the night, suggesting (hopefully) that it isn’t TOO large. While as "flat" as possible a leak line is certainly desirable, it is not necessarily an essential goal. Getting a firm fix on the ABSOLUTE unintentional leak, using the Respironics nightly graph, is not a simple task. Again, more often, a user will simply decide that as long as the leak is largely under, say, 65 lpm (about 40 lpm as a working average for venting, and about 24 lpm as an “acceptable” unintentional leak), everything is probably OK. Some users consider up to 75 lpm an acceptable leak. And, probably, everything is OK. Also helpful and important is the Respironics “Large Leak” graphic indicator, a thick bar telling the user (who may have otherwise been unaware of any potential problem) that the machinery is unhappy with the extent of leak. These “Large Leak” indicators need to be taken seriously, even if we don’t know precisely what criteria are used when they are displayed. The Encore software Help File simply says, “If the chart shows leakage at the “LL” level (large leak), this is an excessive leak that will compromise therapy. This could be the result of a poor mask fitting.”
On the other hand, since Resmed ResScan nightly graphs display only net unintentional leak, the baseline leak rate could be (and ideally will be) zero. A constant, higher baseline suggests a systematic leak, likely due to a mechanical defect or an incorrectly assembled mask fixture. Not having a “Large Leak” indicator, the Resmed software places a horizontal red line at 0.40 lps (= 24 lpm), suggesting that if the leak graph exceeds that level for any significant period of time, leak may be excessive and may require remedial action.
Neither software package is consistently and markedly “better” than the other in all respects. What is important is that the user learn to understand and interpret what the software being used can and cannot, does and does not, tell you, and when, and how, to respond to the information. In any case, the software VASTLY increases the amount of information regarding leak (and other important CPAP parameters) available to the user. If you have a data-capable machine, and if the additional cost of software and associated data-transfer hardware is possible, this is a highly recommended investment.
What Can (And Should) I Do About Leaking?
OK, so you have leak. Well, we all have leak. The most important thing is to keep it in perspective. Not every leak is a PROBLEM leak. Don’t allow yourself to be bullied by the pundits who proclaim, “The first thing you need to do is to get your leak under control.” This is often the first comment of those with nothing better to say, and not enough sense to keep quiet. They usually have no objective or consistent standards on which to base their rants. Still, it is possible that they are right!
ResMed provides a good perspective for thinking about leak. Some of their machines offer a “Mask Fitting Feature” (which is a good place to start if you happen to have such a machine). They clearly indicate that leak is not simply a "good vs. bad" dichotomy, but a spectrum. Some degree of leak is quite acceptable:
Star Rating Description leak l/s leak l/m * * * * * Excellent 0.00 - 0.18 0.0 - 10.8 * * * * _ Very good 0.19 - 0.26 10.9 - 15.9 * * * _ _ Good 0.27 - 0.34 15.7 - 20.4 * * _ _ _ Adjust mask 0.35 - 0.41 20.5 - 24.6 * _ _ _ _ Adjust mask 0.42 - 0.49 24.7 - 29.9 _ _ _ _ _ Adjust mask ≥ 0.50 ≥ 30
The place to start is with selection of a mask. There are several different types (nasal mask, nasal pillows, full-face mask, etc.) and lots of different brands. There are few up-front criteria for making a first decision. You’ll likely be pushed one way or the other by your prescriber or distributor. Not much opportunity or basis for taking control at this point. But what you can and should insist upon is a professional FITTING. Whatever the type and brand of mask, the supplier should assure that you are getting the CORRECT SIZE mask, and that the straps are properly adjusted. Ideally, the mask should be tested with the machine running, to assure a proper fit.
Adjustment of mask straps is not a one-time event. What works well while you’re sitting up is not necessarily what will work well when you’re lying down and moving around. Strap adjustment is always a compromise between minimizing leak and maximizing comfort. Additionally, over time, with use and with washing, the length and elasticity of straps change. Set-it-and-forget-it may work for a while, but periodic readjustment of the straps will likely be necessary.
Typically, the strap lengths on right and left should be symmetrical, to provide even pressure on each side. On the other hand, few of our faces are perfectly formed, and we often have sleeping positional preferences which might make an off-center strap configuration a better choice. Similarly, a fraction of an inch higher or lower may improve a mask’s leak characteristics significantly.
Different masks may require different strap-tightening techniques. Some masks (e.g. ResMed's Activa) need to be set up very loosely - "ballooning" under air pressure provides the seal. Other masks simply have to be strapped fairly firmly. Read the instructions which come with each mask! And remember that when an adjustable angle or length between the headpad and the body of the mask is provided, experimentation to find an ideal positioning to minimize leak may well be worth the time and effort. Again, read the instructions!
Sometimes, the straps supplied with a mask may simply not be adequate. Additional supportive material (ranging from elastic bandages to panty-hose) to hold the mask in place, have been recommended and found effective.
Masks can be very durable, but they aren’t designed to last forever. Periodically check your cushions for tears, or changes in texture and flexibility which can cause increased leaking. Replace them when necessary.
Similarly, check other equipment – particularly the condition of the rubber cuffs on the ends of tubing. And when you reassemble a mask after cleaning, check that all parts fit smoothly and without leak. Some masks have been reported to have “design flaws” which result in excessive leak, which can be easily corrected with such items as Teflon plumbing tape.
Facial oils can be a factor in mask seal leak. Washing the face with soap prior to applying the mask, and/or using an alcohol wipe on the facial skin areas where the mask will be in contact, may prove helpful. Similarly, oils should be gently cleaned off the cushion in the morning after use. (Note that most mask manufacturers do NOT recommend using alcohol directly on mask parts.)
Facial hair can be a challenge in obtaining a good mask seal with minimal leak. Some masks have been reported to present less leak problems than others, with mustaches and beards. Shaving before bed may also help to obtain a better seal.
Various substances have been recommended as “gaskets” to help achieve a better seal. Procedures can be found for using Macks Silicone Ear Plugs (as well as other substances) to build a more air-tight facial seal around cushions. For users of nasal pillows, a variety of commercial products have been suggested for reducing leak as well as enhancing comfort.
Some CPAP users find that “special” pillows, with unique shapes or custom filling material, prevent displacement of their masks and consequent leakage. Others find that a conventional pillow “stabilizes” a mask against their face, preventing leak.
The most effective method of dealing with mouth leak is switching to a full-face mask, which encompasses both the nose and mouth inside the mask. With no pressure differential between mouth and nose, there's no added stimulus to exhale through the mouth. Most importantly, airflow through the mouth (while it still may result in dryness) does not impact the effectiveness of CPAP therapy on sleep apnea. However, there is a price to pay. The much larger cushion needed to surround both the mouth and nose greatly increases the area of skin contact, and thus increases the potential for seal leak. And there are sometimes comfort issues: straps often need to be tighter, irritation of the nasal bridge can become a concern, the masks are often larger and heavier, and covering both the mouth and nose is psychologically uncomfortable for some people. Still, a full-face mask is the "gold standard" solution to mouth leak problems.
Other methods of dealing with mouth leak are widely debated. Many commercial “chin straps” or home-made devices will reportedly help hold the jaw in place so that the mouth does not fall open. Reports of effectiveness of chin straps are mixed. Many varieties of tape, and different taping techniques, have been recommended to hold the lips closed. That lip-taping will reduce or eliminate mouth leak is clear. The absolute safety of this practice is less clear, though the theoretical dangers involved tend to be greatly overstated. There are also indications that for some individuals, lip-taping may paradoxically worsen the efficacy of CPAP therapy. “Training” for the tongue to reduce leak has been suggested, though there is precious little good published science available in this area. There is now a product called the "DeepSleep" made by Goldstream CPAP which can also help prevent airflow out the mouth. It is placed in front of the teeth and behind the lips at bedtime and forms a gasket against the inside of the lips. Many find that it can work when a chin strap does not.
Additional details about the suggestions mentioned above for dealing with leak will be easily found by suitable searches of the forum archives at cpaptalk dot com.
In conclusion, remember that leak is not a problem unless it is a problem. Otherwise, it’s just another four-letter word.
--Velbor 16:24, 10 November 2009 (UTC)
--Velbor 23:53, 19 November 2009 (UTC) edited by Velbor
Mouth Breathing and Mouth Leaks
Mouth breathing and mouth leaks are closely related, but not necessarily the same thing.
Mouth breathing must occur when the nasal passages are blocked, so the only way for any air to get into the lungs is through the mouth. People with blocked noses breathe through their mouth, not through their nose. The solution to mouth breathing is to use a full face mask. Some effort and time may be required to find the best full face mask, fit it, and adjust to it. Use of nasal irrigation and humidified CPAP, especially with nasal pillows, may open nasal passages that were previously chronically blocked. The person can then learn to breathe through the nose.
Mouth leaks may occur if the person is breathing through their nose, but opens their mouth during sleep. If leakage through the mouth is a persistent problem, CPAP treatment can be rendered completely ineffective. CPAP air detouring out the mouth does nothing to keep the airway open. Either mouth breathing or mouth leakage can also cause feelings of suffocation from apneas still happening and/or just the choking feeling of air rushing out the mouth unexpectedly. The safest solution to mouth leakage is to use a full face mask.
Other remedies including using a homemade or commercial chinstrap in hopes that the tongue will maintain an airtight seal inside the mouth if the jaw is kept up. Many people find that chinstraps don’t work to prevent mouth leakage.
Another remedy is to use the tongue to maintain an airtight seal inside the mouth. Some people train the tongue by positioning the tip of the tongue behind the upper front teeth or on the roof of the mouth, and let the tongue spread out in back to seal the throat air passage, even if the lips open. Others use a dental splint, custom made by a dentist, or a do-it-yourself mouth guard to help the tongue maintain an airtight seal.
“The safety of taping the mouth shut has not been proven and there are potential risks of regurgitation and aspiration of food and of suffocation.” TS Johnson MD et al, Sleep Apnea – The Phantom of the Night, p. 167. Mouth taping is especially dangerous for anyone who ever gets blocked nasal passages during the night. If air can’t get in through the nose, it needs to get in through the mouth. Mouth taping is also risky in case of a hose disconnect or power outage.
Warning against mouth leakage
Whether you use a nasal mask or nasal pillows (or any other kind of mask except a full face), if air leaks out through your mouth, the PAP therapy will not work. There are two safe options. The first is to learn to keep your mouth closed while sleeping. Since the feel of pressurized air exiting your mouth is an unpleasant sensation, PAP aids in this. Some people position the tongue behind the top teeth and let it spread out in back to cover the throat opening and make a seal so the PAP works. The second option is to use a full face mask. If you are a mouth breather, breathing in through your mouth instead of your nose, a full face mask is required. An alternative practice for mouth leaks, mouth taping or sealing, is understandable but not advisable, and is not safe if your nose gets stuffy at night, you have acid reflux, need to regurgitate; or you have a hose disconnect or lose machine power and need to breathe through your mouth.
See also Mask Leaks Blow onto Partner
Comment added by --Velbor 21:57, 22 November 2009 (UTC) : The above statement that "if air leaks out through your mouth, the PAP therapy will not work" is far to rigid. As described in the "Introduction to Leak" segment of this article, there are different types and severities of mouth leak. The following tracing (total time 7:09) involving use of a ResMed Activa nasal mask
provides an example of an acceptable pattern of oral exhalation, sometimes referred to as "lip flutter" or "lip puffing", in which the jaw remains closed but the lips open with each exhalation as if blowing a bubble. AutoPAP is set at 10 - 20cm, with median pressure 10.8, 95% pressure 11.4, and maximum pressure 11.8. Note the periodicity of the leak and the return to zero baseline between episodes. This is presumably related to the cyclic variation in sleep stage or depth affecting orofacial musculature; note the latency after each brief arising before these patterns begin. Median leak for the night is 0.00 lps!! Seal leak is negligible. While maximum leak is 0.54 lps, above the "red line", the 95% leak is 0.38 lps, which indicates that the time spent above the 0.40 lps "alert" level is LESS THAN 5% of the night. Efficacy of treatment is excellent (particularly by ResMed standards): AHI = 5.3, AI = 0.0 and HI = 5.3 . While mouth leak is never desirable, care must be taken to avoid arbitrarily labeling it as terrible.