Help!!! Clueless about Flow Limitations

AMUW wrote:SWS, looks like you caught my trap on discussing inhalation vs. exhalation flow shapes.
Which I set unintentionally, and somewhat by insufficient immersion into the breathing process ... and of disordered breathing during sleep. Note that DreamDiver was also highlighting the lower portions of the breathing curve ... that's exhalation I believe.
I am told that in OSA the exhalation phase is more important than the inhalation phase. Not only is CO2 exhaled, but the remaining O2 in the blood stream is extracted. And there is a tight coupling between the gas exchange process and the vaso-constriction in the nose and the pressure variations in the upper respiratory system.
I admit that I haven't looked yet at what ResMed does with the shape of a single cycle, whether the inhalatory or exhalatory half of it. But note that ResMed measures the flow in the box ... not in one's throat or nose.

Sounds good... But again, my understanding is that PSG clinical practice and all APAP algorithms specifically deal with inspiratory flow limitation versus highly uncommon expiratory flow limitation.

I would also add that your document, and others, show "expiratory flow limitation" occurring as flattening specifically during exhalation's end phase. So the exhale part DreamDiver has highlighted in the chair at left is likely not expiratory flow limitation---since it sits at the beginning of exhalation. And the S9 properly refrained from scoring that event as FL. The "M" shape at right occurs exclusively during inspiration and was scored as FL. I doubt those red expiratory tracings in DreamDiver's graph above are comparatively-rare expiratory flow limitations, but I could be very wrong.

Regarding the pneumotach sensor being in the box... That's really not an impediment toward differentiating either flow direction or FL waveshape. The pneumotach is sensitive enough to pick up subtle heartbeat pulsations in the patient flow signal. The sensor location inside the box adds propagation delay to the flow signal compared to proximal sensor at the mask---or even hypothetical flow sensor inside the human airway.

AMUW wrote: And there is a tight coupling between the gas exchange process and the vaso-constriction in the nose and the pressure variations in the upper respiratory system.

I would love to learn more about that resulting vaso-constriction process in the nose should you or anyone have links, AMUW. I have no familiarity with that mechanism and I suspect it's important to UARS and FL patients.


Again, thanks for posting all that, AMUW... Very good discussion, indeed.

-SWS wrote:Sounds good... But again, my understanding is that PSG clinical practice and all APAP algorithms specifically deal with inspiratory flow limitation versus highly uncommon expiratory flow limitation.

"IMHO", I believe that EFL is quite common, but that it's origin cannot be specifically identified using only flow technology.

Consequently, one could easily end up with a pressure assault against an asthma attack.

Jnk's advice about trending can apply to flow limitations as well. Here's an example of taking advantage of ResScan's S9 scoring for purposes of trending FL at home:

Below we have ResScan FL scoring using 4cm CPAP on 4-5-11 and 5cm CPAP on 4-7-11:


I just performed a screen grab of SU's post above, and cropped last night's 6cm-to-8cm APAP for comparison:

Bright Choice wrote: Can any of you direct me to where I could look at some "normal" flow graphs and some "abnormal" flow graphs so I could begin to understand what I am looking for.

Here are a few examples of normal, rounded flow curves (top), versus flow curves with flow limitation and snoring superimposed:
http://chestjournal.chestpubs.org/conte ... .large.jpg
My understanding is that flow-curve "morphology" across the SDB patient population is spread across a wide continuum----let alone often with hybridized SDB flow-shape characteristics. So the above examples wouldn't necessarily translate to your own flow-limited waveshape characteristics. That's one reason taking advantage of the S9 ResScan FL graph for purposes of trending, as above, might make more sense as jnk already mentioned.

Bright Choice wrote: I think I am dealing with UARS but have some question about the "centrals".

Different doctors seem to have different views and definitions of UARS. Dr. Guilleminault, the father of UARS, views that UARS and OSA are mutually exclusive---with UARS being an airway hypersensitivity disorder and OSA being an airway hyposensitivity disorder. So if you went to his Stanford sleep clinic and were diagnosed with OSA, then he would not also diagnose you with UARS. Others, like Dr. Krakow, claim that UARS and OSA are not mutually exclusive.

NotMuffy wrote:

-SWS wrote:Sounds good... But again, my understanding is that PSG clinical practice and all APAP algorithms specifically deal with inspiratory flow limitation versus highly uncommon expiratory flow limitation.

"IMHO", I believe that EFL is quite common, but that it's origin cannot be specifically identified using only flow technology.

Consequently, one could easily end up with a pressure assault against an asthma attack.

Thanks for that, NotMuffy... Your experienced IMHO's work well for me...

Here are interesting FOT-derived impedance calculations (top two curves):

http://www.scielo.br/scielo.php?script= ... 9000700008

The pair of curves at left are the control group, and the pair of curves at right are asthmatics (note different scales). I'm not sure what that curve comparison can tell us with respect to sleep disruption. Those are wake flow curves rather than sleep flow curves, so I expect asthmatic sleep flow curves can look very different...

But the airway impedance calculations suggest that the asthmatics will sometimes require significantly more respiratory effort during sleep. And that, in turn, suggests that asthmatics are probably more susceptible to sleep related RERAs than "normals'.

-SWS wrote:Jnk's advice about trending can apply to flow limitations as well. Here's an example of taking advantage of ResScan's S9 scoring for purposes of trending FL at home

Although I haven't yet used an S9, I would assume that the flow-limitation data would be, for many, the particularly/especially useful data for trending with that machine. Is that how you see it?

I also assume that the answers to the questions, "Am I having (some) flow limitations?" and, "Am I having one or two (natural) central events?" are very often "yes" and "yes," but that those questions/answers give less useful treatment-tweaking information than the questions, "How does my data (including FL data) for the most recent two weeks compare to my data for the preceding two weeks?" and, "Is there any pressure adjustment I could make that might improve the data (including FL data) and improve how I feel?"

But that viewpoint of mine may be based mostly on my personal experience of believing that my initial diagnosis was right on the money and my approach as a self-tweaker of my pressures.

Do expiratory flow limitations cause symptoms?

jnk wrote:

-SWS wrote:Jnk's advice about trending can apply to flow limitations as well. Here's an example of taking advantage of ResScan's S9 scoring for purposes of trending FL at home

Although I haven't yet used an S9, I would assume that the flow-limitation data would be, for many, the particularly/especially useful data for trending with that machine. Is that how you see it?

I suspect heuristically trending that S9 flow-limitation graph, as in the example above, will be a more friendly analysis technique for most people.

jnk wrote: I also assume that the answers to the questions, "Am I having (some) flow limitations?" and, "Am I having one or two (natural) central events?" are very often "yes" and "yes," but that those questions/answers give less useful treatment-tweaking information than the questions, "How does my data (including FL data) for the most recent two weeks compare to my data for the preceding two weeks?" and, "Is there any pressure adjustment I could make that might improve the data (including FL data) and improve how I feel?"

Agreed, since most of us have basic OSA. But in her first post, Bright Choice tells us that those treatment parameters are already accounted for. If she has outstanding symptomology, then looking for possible FL or central tendencies in home data is pretty much what this message board does well. So you correctly told Bright Choice the best first-place to look for problems IMO. But it sounds as if she already has her frank obstructive events under control.

Rhetorically: in Bright Choice's situation should a patient continue to scrutinize their basic obstructive events some more, despite good obstructive AHI, or is time spent exploring FL and CA a worthwhile expenditure?

jnk wrote: But that viewpoint of mine may be based mostly on my personal experience of believing that my initial diagnosis was right on the money and my approach as a self-tweaker of my pressures.

Many posters here seem to devote significant time analyzing their data---toward validating or invalidating their beliefs and assumptions. I can't see that being a bad thing or waste of time, unless that data-analysis pursuit fosters anxiety. I tend to look at my own data every couple/few months.

Thanks. Good points, -SWS, as always. I appreciate it.

To me, trending with FLs makes equally as much sense as trending with AHI and snores. Techs try to titrate out snores and less-than-hypopneas when they can, I understand, so it makes good sense to me that we plain-OSA self-tweakers do the same before assuming our symptoms are due to something besides the bad breathing that relates to our primary diagnosis. So I am glad the OP asked the question and that you are helping so many of us gain a better understanding of flow limitations.

Bright Choice wrote:I think I am dealing with UARS but have some question about the "centrals".

What questions would you like to ask, Bright Choice? Jnk's comments about how SDB learning time might best be spent was certainly not meant to dissuade you---or anyone---from asking whatever SDB questions you would like to better understand.

So fire away with your questions about flow limitations or central apneas and people here will try their best to answer your questions.

AMUW wrote: While I expect that an experienced sleep doctor can recognize a lot in the time histories [and their trends over successive nights], even better derived measures have been proposed (if the doctor and tech can handle the higher signal processing math and physics); for example the ratio of pressure to flow, plotted vs. time estimates the upper airway resistance... or frequency spectra ... or random analysis techniques.

NotMuffy wrote:I believe that EFL is quite common, but that it's origin cannot be specifically identified using only flow technology. Consequently, one could easily end up with a pressure assault against an asthma attack.

-SWS wrote: Here are some interesting ... airway impedance [derived curves]

Wow, you two folks are ahead of me ... I'll try to catch up:
1. I used the simpler term "resistance" above, since it stuck in the UARS medical terminology.
[Edited later: sorry, I didn't see the link to the Brazilian paper; still wonder about the meaning of averaging time curves]; if they mean complex impedance between two periodic variables (not real amplitude ratios), then I wonder whether the phase difference makes sense ... at the same location in the patient-on-CPAP system.
[added later]: Do the first 2 curves in Fig.1 (assuming that the phasing is correct) indicate that the airway resistance is worst during the expiratory phase of that 4-5 sec breathing cycle? Also, that's 20-25 breaths per minute; isn't that rather high for an awake adult? or are they children?

Talking about "smarter" derived quantities than basic statistics (average and 95-percentile) and visual description of higher order variations... has any of you considered adopting the spectral and random analysis techniques from other signal processing disciplines (including cardiology). Extract from one reference:

Multivariate Analysis of Blood O2 Satn Recordings in OSA dx. Alvarez (Univ.Valladolid, Spain) IEEE Trans. Biomedl Engg, Dec.2010 Abstract: This study focuses on the analysis of blood oxygen saturation ... 148 patients suspected of suffering from OSA ... Our feature set included common statistics in the time and frequency domains, conventional spectral characteristics from the power spectral density (PSD) function, and nonlinear features... We conclude that simultaneous analysis in the time and frequency domains by means of statistical moments, spectral and nonlinear features could provide complementary information ... to improve OSA diagnosis.

2. I like your periodic breathing curves showing snoring in either inspiration or exhalation. Which, when the snoring disorder progresses to full closure, means to me that apneas are just as possible on either half of the breathing cycle ... that what really counts is the pressure difference across the pharynx

3. SWS, would you have an explanation or illustration on how you use the FL history to link to breathing disorder ... instantaneous use or as a trend over time to estimate RERAs? My naive way, maybe similar to the S9: I see a precursor evidence to an apnea ... so it is my understanding that the S9 Auto algorithm bumps up the pressure. If one's flow limitations arrive in clusters -- OSAs, or maybe UARS, but not Cheyne-Stokes periodic breathing -- then the Auto pressure can quickly end up at the upper limit and cause mask leakage; once resolved, the pressure dies off asymptotically to the lower limit

4. JNK, could you define your use of "trending" in this context?

jnk wrote: To me, trending with FLs makes equally as much sense as trending with AHI and snores. Techs try to titrate out snores and less-than-hypopneas when they can ... you are helping so many of us gain a better understanding of flow limitations.

AMUW wrote:. . . 4. JNK, could you define your use of "trending" in this context?

jnk wrote: To me, trending with FLs makes equally as much sense as trending with AHI and snores. Techs try to titrate out snores and less-than-hypopneas when they can ... you are helping so many of us gain a better understanding of flow limitations.

Sorry. I was tossing around slang. I misused the word "trending" to mean 'watching the trends in my home-machine data." I get sloppy with my wording like that at times.

What I and a few others do for fun and comfort and all-around health is to try to find the best range of pressures (or a single CPAP pressure, if not using an auto) for ourselves by choosing our minimum pressure (or single pressure, in not using an auto) based on how the data from our machines trend when we adjust the pressure. I was taught how to do that here in this forum.

My version of that method was that I kept raising my minimum by a cm for a week or two until I no longer noticed any improvement in my data or how I felt.

What I was trying to say was that if I were using an S9 Autoset, I would probably keep raising my minimum until there was no improvement in my AHI or in my flow limitation charts. If raising pressure didn't improve my AHI, my charts, or how I felt, I would undo the change. The idea is to use the pressure, or range of pressures, that most comfortably provides the best breathing and sleep.

I very much appreciate your pointing out my misuse of the word, AMUW.

AMUW wrote: I used the simpler term "resistance" above, since it stuck in the UARS medical terminology. SWS, you didn't provide the full reference, but if you mean complex impedance between two periodic variables (not real amplitude ratios), then I wonder whether the phase difference is being used ... at the same location in in the patient-on-CPAP system.

Here's the link that goes with the asthmatic graphs, AMUW: http://www.scielo.br/scielo.php?script= ... 9000700008
The above technique and S9 CPAP both employ versions of monofrequency Forced Oscillation Technique (mFOT). They both introduce low-amplitude, single-frequency pressure oscillations to the airway----and measure pressure/flow phase difference during spontaneous respiration---to derive complex impedance, including an imaginary component as airway reactance. But other than the multi- and monofrequency Forced Oscillation Techniques, I think the term airway "resistance" (without the imaginary component) is most often used in medical literature.

AMUW wrote: Has any of you considered adopting the spectral and random analysis techniques from other signal processing disciplines (including cardiology).

Bear in mind I adopt a mere message board technique, since I have no link with any of this other than as a patient. But sleep medicine research relies fairly heavily on those techniques. Cyclic Alternating Pattern (1st two links below) is a notable example:
http://scholar.google.com/scholar?hl=en ... =&as_vis=0
http://www.medscape.com/viewarticle/494651_5
http://www.clinph-journal.com/article/S ... 6/abstract

AMUW wrote: I like your periodic breathing curves showing snoring in either inspiration or exhalation. Which, when the snoring disorder progresses to full closure, means to me that apneas are just as possible on either half of the breathing cycle ... that what really counts is the pressure difference across the pharynx

Expiratory end-phase is when pressure across the pharynx starts to wane I believe---and that's when obstructive problems often become incipient. Accordingly, BiLevel titration protocol typically addresses obstructive apneas with EPAP pressure increases.

AMUW wrote: SWS, would you have an explanation or illustration on how you use the FL history to link to breathing disorder ... instantaneous use or as a trend over time to estimate RERAs? My naive way, maybe similar to the S9: I see a precursor evidence to an apnea ... so it is my understanding that the S9 Auto algorithm bumps up the pressure. If one's flow limitations arrive in clusters -- OSAs, or maybe UARS, but not Cheyne-Stokes periodic breathing -- then the Auto pressure can quickly end up at the upper limit and cause mask leakage; once resolved, the pressure dies off asymptotically to the lower limit

I think sleep medicine has viewed treating FL in these two contexts: 1) indicator of initial airway patency problems---and thus precursor to more severe obstructive apneas and hypopneas, and 2) a potential disturbance mechanism irrespective of apneas and hypopneas (RERA based disturbances). I don't currently have the trend or FL/RERA correlative type data you mentioned. I think this presentation by Resmed gives a nice illustration/explanation of FL leading to RERA:
http://www.resmed.com/us/multimedia/und ... 40x380.swf

And this 2009 presentation by Dr. Guilleminault is replete with useful information IMO:
http://www.ucsfcme.com/2010/slides/MOT1 ... tIsNew.pdf

jnk wrote: I very much appreciate your pointing out my misuse of the word [trending]

One can numerically trend or heuristically trend. Visual comparison of the FL graphs above is an example of heuristic trending rather than numeric trending. So your word choice was just perfect IMHO, Jeff.

AMUW wrote: Has any of you considered adopting the spectral and random analysis techniques from other signal processing disciplines (including cardiology).

I use R-R Interval to analyze Heart Rate Variability (HRV) grossly; Kubios to look locally at HRV, but exclusively as a teaching tool; review ECG-Derived Respiration (EDR) reports from a referral source; and although thought at one time would get the Cardiopulmonary Coupling (CPC) Module (which combines HRV and EDR) after the buyout (presently available in Remlogic), it appears that availability was only a rumor.

CPC is slick:



but keep in mind that the use of these routines is largely limited to diagnosis, and only look impressive in the face of significant disease when reviewing records in their entirety. "IMHO', once people get to the Board, the macroanalysis used in spectrofourniomumbojumbo (attempting to reduce complex signals to a "simple" result) is not helpful and needs to be unbundled.

-SWS wrote:. . . Rhetorically: in Bright Choice's situation should a patient continue to scrutinize their basic obstructive events some more, despite good obstructive AHI, or is time spent exploring FL and CA a worthwhile expenditure? . . .

Am I allowed to comment on rhetorical questions?

As a patient, I simplistically tend to think of all the numbers and charts from home machines as information that clinicians and patients can use over time to track the stability of the upper airway in response to pressure. I tend to prioritize in my mind the significance of the reported events into a simple hierachy in which apneas are worse than hypopneas and hypopneas are worse than flow limitations and snores.

Therefore, if central apneas were to increase in response to my chasing flow limitations and snores, I would see that as an increase in AI that didn't go away, and I would, of course, lower my pressure, since apneas trump flow limitations and snores in my simple approach.

That is why, to my mind, the self-tweaking approach is self-correcting when it comes to flow limitations and central apneas, in the context of pressure choices made by a self-tweaker for treatment purposes.

So I tend to think in terms of "what is the best pressure," since, as I see things, it would be only after the best pressure possible has been tried with CPAP/APAP and there is still insufficient response to treatment that other kinds of machines and treatment approaches should come into play. I don't only want to know if reasonably-successful PAP therapy solves the sleep problems; I want to know if customized, optimal PAP therapy solves the sleep problem. If not, then it is time to try something else beyond simple PAP, I think.

I realize, though, that my overly-simplistic approach might not turn out so well for some people with more complicated conditions than mine (such as hypersensitivities). My words could tend to sound like I'm blaming the victim for not optimizing his own therapy. That is never my intent. I am attempting to be of help to those patients who say "I use the machine and my AHI is reported to be below 5 and I still feel like crap" by saying in response "try to get your treatment optimized by self-tweaking before you give up on PAP, but if that doesn't work, then there are other things to try."

In the meantime, a few centrals here and there and some flow limitations may be part of the process--but part of the process that should be self-correcting when paying attention to trending data, both numerical (AHI) and visual (FL charts).

I hope I am not misleading anyone when I pass on that approach.

Does that sound like a reasonable "first responder" approach?

jnk wrote: Does that sound like a reasonable "first responder" approach?

I think your answer about how newcomers might best spend their time is fantastic in the right context, Jeff. One example of a spot-on context, IMO, would be if a newcomer specifically asks how they might best spend their learning/troubleshooting time. However, the blue text below was the original poster's only central-apnea comment in the entire thread:

Bright Choice wrote: I think I am dealing with UARS but have some question about the "centrals".

BTW, what exactly were the original poster's central apnea questions? The message board never had a chance to hear them, let alone weigh her concerns. Rather, her question was sent to the time-poorly-spent pile based on SDB assumptions and generalizations.

On whole, I'd prefer SDB and CPAP questions not be dismissed before they are asked. Bear in mind I'm an autonomy freak as well: I think our newcomer adults should be encouraged to think, troubleshoot, and manage their own health care concerns----with us as their advisors rather than managers. But hey, that's just my opinion...