Deep Stops Increases DCS

[rossh]

I don't want to come across as a contrarian here, but the idea of emboli in my blood scares the crap out of me. Isn't the very definition of an embolism an occlusion?

In addition, I was taught and have taught my students that the incidence of DCS is tantamount to noise, statistically speaking. In that regard, an incidence of %4 is way above statistical noise and that also scares the crap out of me.

Do these figures apply ONLY to decompression diving? Only to decompression diving under certain deco programs? I've done way over a hundred deco dives, how have I escaped DCS for so long? I feel like you're telling me I've been playing Russian Roulette for a long time.

All this applies to NDL dives as well.

VGE has been running around our system ever since we got in the water. Our NDL dives have VGE. Half of all recreational dives have VGE. We can get VGE from exercise in the gym ! Predive vibrations can shake them loose and we get less, and a hot bath makes them deminish too. They are with us always.

more papers:

Flying after diving - recreational week; Flying after diving: in-flight echocardiography after a scuba diving week. - PubMed - NCBI
https://www.uhms.org/images/DCS-and-AGE-Journal-Watch/germonpr__pre-dive_vibration.pdf
Sauna bath: Predive sauna and venous gas bubbles upon decompression from 400 kPa. - PubMed - NCBI

It goes on and on. Don't be afraid of VGE, or people trying to scare us about them.. We can't avoid them, and they run around our system harmlessly for 99.99% of the time..... Of course, if an individual has a history of skin issues (Like the DAN PFO workshop says), then they probably should not be engaging in high stress diving like deco diving, anyway.

We can certainly lower our tissue pressure stress levels in diving by making profiles longer and slower, and the VGE count will come down in a relative way, but its not the change in VGE that matters, because the were not doing any harm in the first place.

 

[rossh]

Which begs the obvious question:

For the most basic case of a diver with a stable and fixed-size PFO, would it not make more sense to time penalize the first few fast compartments rather than adding conservatism to all of the compartments?

It's a slower tissue problem, and the person with a PFO needs to do more deco, in order to compensate for the shortfall in off gassing rates. I'm still undecided about how to handle this.

 

[David Doolette]

As they say in the paper, "This illustrates that having detectable bubbles yields only a 4% chance of developing DCS. Even when using high bubble grades as test criterion, the positive predictive value is low..."
***********
The simple reality: Fast deco has high supersaturation tissue pressure (decompression) stress, fast offgas rates, and it excites VGE into higher growth. BUT... the reverse implied connection is wrong. High VGE does not imply high decompression stress, because high VGE can / do occur even when only lower decompression stress exists. The correlation is not 1:1.

This seems like groundhog day. This description contains the same confabulation of two related, but different, possible uses of VGE measurements, that was clarified on the Rebreatherworld thread three years ago.

Deep stops debate (split from ascent rate thread) - Page 86

The distinction between these possible uses of VGE measurements is important, particularly as some of the recent posts indicate people are making such measurements following their own diving.

As some background, VGE are graded using ordinal scales typically ranging from 0 to IV, that correspond to the number of bubbles heard in Doppler blood flow signals in large veins or seen in 2-D echocardiographic images of the heart chambers. These VGE grades have a general correlation with the incidence of DCS in large compilations of data. The biggest of these compilations is from the development of the Canadian Forces decompression tables and some of these data have appeared in several places, including Bennett and Elliott’s Physiology of Medicine and Diving, and were analyzed by David Sawatzky in his M.Sc thesis. If you separate the dives into those in which the maximum grade detected following diving was 0 or I or II etc, there is an increasing incidence of decompression sickness (DCS) in the successive higher VGE grades. For instance in air diving: grade 0 (0 DCS/ 819 dives); grades I (3 DCS/287 dives); grade II (2 DCS/183 dives; grades III (27 DCS/365 dives); and grade IV (9 DCS/72 dives). I do not have heliox and trimix data in front of me, but it contains the same correlation. This correlation arises in part because VGE can cause some manifestations of DCS, but also, because VGE form in response to the same stresses that form bubbles elsewhere in the body, VGE are probably correlated with an increase risk of bubble formation at other (extravascular) DCS site.

So, on to the possible uses of VGE.

1) Can you use the VGE grade in an individual diver to diagnose DCS or predict if they will develop DCS? No, the relationship is not strong enough. This is use for which the poor “positive predictive power” of VGE, which has been mentioned, is relevant. To illustrate this, let us suppose that we decided to diagnose DCS in divers on the basis of a VGE score of IV, i.e. any diver who shows a score VGE grade of IV gets tossed in the chamber and treated whether they have symptoms or not. The data above suggests that we would be correct about 13% of the time and wrong about 87% of the time. These latter 87% ‘false positives’, without going into details of the calculations, are the essence of poor positive predictive power.

2) Do the VGE grades measured in a group of divers, all who have performed the same dive profile, provide an estimate of risk of DCS of that decompression schedule? Yes. To illustrate, instead of an individual diver, imagine 50 divers conducted an identical dive profile (dive profile 1) and all 50 divers manifest VGE grade I and another 50 divers conducted dive profile 2 and all 50 divers had VGE grade III. We still would not know if any particular diver would go on to develop DCS, but we could estimate, with considerable confidence based on the data above, that profile 2 had a greater risk of DCS than profile 1, and even (with less confidence), that dive profiles 1 and 2 had risks of DCS of about 1% and 7%. This is an oversimplification, but again this is the essence of the use of VGE to estimate the risk of DCS.

2a) Are VGE used to validate decompression schedules? Yes. The gold standard for validating decompression schedules – and the method still used by the U.S. Navy - is the incidence of DCS, i.e. you test a schedule and accept it if it results in a low incidence of DCS and reject it if it results in a high incidence of DCS. However, many other navies now use VGE as well as, or instead of, DCS. The best example is the validation of the Canadian Forces (DCIEM) decompression tables, which produced the data mentioned above. In this approach, the maximum VGE measured after diving is considered a measure of “decompression stress”, and a schedule is accepted if it consistently produces low VGE grades and rejected if it consistently produces high VGE grades. The DCIEM criteria for heliox schedules is to reject a schedule if in 20 dives the median VGE grade is III or higher.

 

[RyanT]

Ross is correct in saying that VGE can be produced on even very moderate dives, and likewise correct that the correlation between VGE and DCS is not 1:1. But...nothing in physiology ever is! While the venous bubbles themselves may be harmless, they were produced from the same over saturation stresses that generate DCS. As David showed, there is a positive proportional relationship between VGE grades and incidences of DCS. Sadly, no one understands why a high VGE load in one person may fail to generate DCS (or conversely a low VGE score generates a hit). As much as I enjoy quibbling over statistics, at the end of the day, I'll choose on the low side of empirical proportions every time.

 

[UWSojourner]

I don't want to come across as a contrarian here, but the idea of emboli in my blood scares the crap out of me. Isn't the very definition of an embolism an occlusion?

In addition, I was taught and have taught my students that the incidence of DCS is tantamount to noise, statistically speaking. In that regard, an incidence of %4 is way above statistical noise and that also scares the crap out of me.

Do these figures apply ONLY to decompression diving? Only to decompression diving under certain deco programs? I've done way over a hundred deco dives, how have I escaped DCS for so long? I feel like you're telling me I've been playing Russian Roulette for a long time.

Hey NetDoc

Ross's "arguments" trying to downplay the importance of VGE as a decompression stress indicator and, more particularly, of studies linking high grade VGE to the use of deep stop models have no merit. He misreads, either purposely or from a lack of background, the articles he encourages everyone to review.

We've obviously entered a strange world when the bubble-model-marketer has to try to convince everyone that bubbles don't matter in order to sell his concept of decompression. But it's the world we're in. The idea itself begs the question of why researchers would continue to study VGE as a topic in decompression research if VGE is so meaningless to the topic of decompression efficacy? Strange idea.

Consider the following:

• High grade VGE is not equal to DCS -- Obviously this is true. Similar statements would be "Smoke in the lungs does not equal lung cancer", "driving intoxicated is not equal to a car accident", or in your example "Playing Russian Roulette is not equal to severe brain injury." In all cases you can't say the first necessarily implies the second, but you CAN say the first indicates a higher risk of the second.

• High grade VGE is indicative of a higher risk of DCS; Zero, or lower grade, VGE is indicative of lower DCS Risk -- Again, in your example, I think it's safe to say that diving that requires decompression stops is, in a sense, like playing Russian Roulette regardless of the model you use. You can't escape risk entirely. But if you dive models that more frequently give you grade 3-4 bubbles, you are playing the game with more bullets in the gun. If you have a gun that can hold 1000 bullets, diving with no detectable VGE loads 1 bullet; diving a decompression algorithm that gives you grade 3-4 VGE might load, for example, 40 bullets into the gun. Either "game" is still a random exercise (i.e. we can't predict whether any particular person playing the game will lose). But a rational diver would prefer models that lower VGE scores (i.e load fewer bullets) all other things being equal.

 

[David Doolette]

By deep stops, I understand this to mean stops significantly deeper than those mandated by a pure Haldanean model (USN, ZH-16 without GF etc) these will vary with any depth profile so can be very hard to put a number on.

I think the main problem here is that the conversation has devolved into an argument over which profiles are safe and which aren't etc etc whereas the main topic before was the efficiency of the different approaches.

My take-away from the NEDU study and all of the threads on here and RBW etc etc is that its the philosophy that is under question, not specific implementations. To paraphrase the 97 pages on this thread so far, and 1000's on others, my understanding so far:

1. In the past, as a result of numerous factors, including Dr Pyle's experiences and the gang at WKPP, a theory arose that there may be significant advantages to stopping deeper than previously the norm. Much of the evidence was based on "feeling better" after similar run-time dives as well as Doppler imaging after dives conducted by, among others, Dr Wienke.

2. This mindset led to the development of several theories to try account for the visible results and so some of the more popular "bubble models" began to appear. These were all designed with different methods but had in common that their profiles tended to have initial stops somewhat deeper than those prescribed by the USN tables etc.

3. Over the years these bubble models have been tweaked and refined to allow different levels of conservatism etc so the profiles have diverged a great deal depending on what the user inputs (VPM+?, GF etc) but all still have deeper stops than USN and Haldanean models (Buhlmann etc) mandate.

4. All these models have been based on theory and mathematical equations, as well as Doppler of divers returning from certain dives, mainly because nobody was volunteering to get bent for the sake of testing. Also, there is a large track record of "safe" dives by which I mean no reported DCS using these algorithms.

5. The US Navy observed all the fuss, and since one of the many claims over the years was that "bubble models get you up cleaner and faster than the old school", decided to experiment with the new (to them) bubble models with attendant deeper stops.

6. In order to design a suitable experiment they decided to keep all variables the same, including temps workload etc. They also designed the experiments to result in a measurable DCS incidence in order to confirm the edges of the envelope. In order to assess the efficiency of deep vs shallow stops, they designed profiles with the same runtime with the stops allocated deeper in one profile (A2) than in the other profile (A1). Their reasoning was that if the deeper stops were more beneficial, they should see a lower incidence of DCS in A2. Doubtless if that had been the case they would have continued the study in order to determine which particular algorithm got their divers out of the water cleanest, fastest.

7. What they found in the initial proof-of concept testing was that the A2 profile was significantly more DCS prone than the model they were currently using so they decided to stop further testing. In trying to decide why this would be (which the navy is not interested in as they won't be going to deep stops) the researchers involved crunched all the data to see if they could isolate a cause. They finally settled on the most likely candidate being that the slower compartments were remaining more saturated for longer and this would lead to more deco stress on the diver after surfacing. This is contrary to the previously prevailing opinion that it is more important to protect the fast tissues than the slow ones, hence deeper stops.

8. Since we are lucky enough to have members of that team on these and other boards, they elected to inform us of the results of that test and their best inferences of how it would be applicable to the diving we do. I know Dr Mitchell has upped his GFLo, leading to shallower first stops for his personal diving so he obviously believes that the test had relevance to him. He also presented a talk entitled "Decompression Controversies" (understatement of the year anyone?) which was disseminated on here and Youtube etc

9. Numerous role players have taken the above as an attack on them and their work (rightly or wrongly) and so the resulting discussion has become like a religious argument, with all the logic and reasoned debate that goes with that.

Me? Ive changed to a higher GFlo now, still deeper stops than a pure Buhlmann but not as deep as previously. I feel no worse than before and my friends who changed have reported at worst feeling no different and at best substantially better afterwards.

I also have many friends diving VPM with varying settings and so far all safe, so there are many ways to skin this particular cat but which one is the most efficient, least messy?

Thanks for the patience, this was almost as much for me to get my head around where we are at on this subject. PLEASE correct my many mistakes, I have no doubt Ive missed a lot.

I think this is a very accurate summary. The one thing I would clarify is in item 1. I am not sure there was much Doppler work done in support of deep stops at the time of their popularization – at least I know of very little published data. One of the key documents often cited in support of deep stops was ”Neuman TS, Hall DA, Linaweaver PG. Gas phase separation during decompression in man: ultrasound monitoring. Undersea Biomed Res 1976;3:121-30.” This paper reported Doppler VGE measurements during decompression in a dry chamber (i.e. during the decompression stops). It is a fascinating paper because there has been relatively little work done on VGE at depth. What they found was that as a result of an unplanned “deep stop” (a short extra stop inserted 10 fsw deeper than the prescribed first stop to allow the watches to catch up after a overly-rapid first pull), there were fewer detected VGE during the decompression stops. This makes sense. However, they did not continue to measure VGE long after surfacing, so may not have detected the maximum VGE following diving – which is the VGE measurement that is correlated with the incidence of DCS (and in more recent work has generally been shown to be increased by deep stops). They did not do enough dives to find a significant difference in DCS, and most subjects got at least skin bends, as it was a very aggressive profile. A rarely cited report of the same era, “Spencer MP, Johanson DC. (Institute of Environmental Medicine and Physiology). Investigation of new principles for human decompression schedules using the Doppler ultrasonic blood detector. Technical Report. Washington DC: Office of Naval Research; 1974 Jul.”, reported a comparison of a shallow stops versus a deep stop schedule, those being a 200 fsw / 30 minute BT schedule from the Dwyer/ des Grange 1957 Standard Air decompression Table (USN Diving Manual from 1959-2008), which has a 40 fsw first stop, and the corresponding schedule from the old C&R tables (derived from Haldane’s), which has a 70 fsw first stop (and six minutes shorter total decompression time). That was a small comparison with just a couple dives on each, but the deep stops resulted in more VGE. As alluded to above, recent work – in the last decade – has indicated that compared to shallow stops, deep stops during decompression dives result in either the same or higher maximum VGE grades following diving.

 

[scubadada]

By deep stops, I understand this to mean stops significantly deeper than those mandated by a pure Haldanean model (USN, ZH-16 without GF etc) these will vary with any depth profile so can be very hard to put a number on.

I think the main problem here is that the conversation has devolved into an argument over which profiles are safe and which aren't etc etc whereas the main topic before was the efficiency of the different approaches.

My take-away from the NEDU study and all of the threads on here and RBW etc etc is that its the philosophy that is under question, not specific implementations. To paraphrase the 97 pages on this thread so far, and 1000's on others, my understanding so far:

1. In the past, as a result of numerous factors, including Dr Pyle's experiences and the gang at WKPP, a theory arose that there may be significant advantages to stopping deeper than previously the norm. Much of the evidence was based on "feeling better" after similar run-time dives as well as Doppler imaging after dives conducted by, among others, Dr Wienke.

2. This mindset led to the development of several theories to try account for the visible results and so some of the more popular "bubble models" began to appear. These were all designed with different methods but had in common that their profiles tended to have initial stops somewhat deeper than those prescribed by the USN tables etc.

3. Over the years these bubble models have been tweaked and refined to allow different levels of conservatism etc so the profiles have diverged a great deal depending on what the user inputs (VPM+?, GF etc) but all still have deeper stops than USN and Haldanean models (Buhlmann etc) mandate.

4. All these models have been based on theory and mathematical equations, as well as Doppler of divers returning from certain dives, mainly because nobody was volunteering to get bent for the sake of testing. Also, there is a large track record of "safe" dives by which I mean no reported DCS using these algorithms.

5. The US Navy observed all the fuss, and since one of the many claims over the years was that "bubble models get you up cleaner and faster than the old school", decided to experiment with the new (to them) bubble models with attendant deeper stops.

6. In order to design a suitable experiment they decided to keep all variables the same, including temps workload etc. They also designed the experiments to result in a measurable DCS incidence in order to confirm the edges of the envelope. In order to assess the efficiency of deep vs shallow stops, they designed profiles with the same runtime with the stops allocated deeper in one profile (A2) than in the other profile (A1). Their reasoning was that if the deeper stops were more beneficial, they should see a lower incidence of DCS in A2. Doubtless if that had been the case they would have continued the study in order to determine which particular algorithm got their divers out of the water cleanest, fastest.

7. What they found in the initial proof-of concept testing was that the A2 profile was significantly more DCS prone than the model they were currently using so they decided to stop further testing. In trying to decide why this would be (which the navy is not interested in as they won't be going to deep stops) the researchers involved crunched all the data to see if they could isolate a cause. They finally settled on the most likely candidate being that the slower compartments were remaining more saturated for longer and this would lead to more deco stress on the diver after surfacing. This is contrary to the previously prevailing opinion that it is more important to protect the fast tissues than the slow ones, hence deeper stops.

8. Since we are lucky enough to have members of that team on these and other boards, they elected to inform us of the results of that test and their best inferences of how it would be applicable to the diving we do. I know Dr Mitchell has upped his GFLo, leading to shallower first stops for his personal diving so he obviously believes that the test had relevance to him. He also presented a talk entitled "Decompression Controversies" (understatement of the year anyone?) which was disseminated on here and Youtube etc

9. Numerous role players have taken the above as an attack on them and their work (rightly or wrongly) and so the resulting discussion has become like a religious argument, with all the logic and reasoned debate that goes with that.

Me? Ive changed to a higher GFlo now, still deeper stops than a pure Buhlmann but not as deep as previously. I feel no worse than before and my friends who changed have reported at worst feeling no different and at best substantially better afterwards.

I also have many friends diving VPM with varying settings and so far all safe, so there are many ways to skin this particular cat but which one is the most efficient, least messy?

Thanks for the patience, this was almost as much for me to get my head around where we are at on this subject. PLEASE correct my many mistakes, I have no doubt Ive missed a lot.

I have found this thread interesting and thought provoking. It has prompted be to do more reading and numerous sample calculations. I still struggle with application and relevance to the majority of divers.

One of the challenges is comparing different algorithms, but with the same run time. Using the 170 foot air dive for 30 minutes, I wonder how VPM +0 would compare ZHL-16C 80/85. They would both have the same decompressiom time of about 81 minutes, but a very different stop distribution, as polar opposite as possible. VPM would start with a very brief stop at 100 feet, whereas ZHL starts with a brief stop at 60 feet. VPM would have 46 minutes between 20 and 10 feet, whereas, this would be 61 minutes with ZHL.

So, nobody would dive these profiles, but I think it illustrates the point. Maybe the profiles we do dive are considerably more conservative than this and the rate of DCS is just very, very low. Maybe the settings we choose make the algorithms more similar than more different.

 

[RainPilot]

I have found this thread interesting and thought provoking. It has prompted be to do more reading and numerous sample calculations. I still struggle with application and relevance to the majority of divers.

One of the challenges is comparing different algorithms, but with the same run time. Using the 170 foot air dive for 30 minutes, I wonder how VPM +0 would compare ZHL-16C 80/85. They would both have the same decompressiom time of about 81 minutes, but a very different stop distribution, as polar opposite as possible. VPM would start with a very brief stop at 100 feet, whereas ZHL starts with a brief stop at 60 feet. VPM would have 46 minutes between 20 and 10 feet, whereas, this would be 61 minutes with ZHL.

So, nobody would dive these profiles, but I think it illustrates the point. Maybe the profiles we do dive are considerably more conservative than this and the rate of DCS is just very, very low. Maybe the settings we choose make the algorithms more similar than more different.

My gut feeling on all of this is that for the mythical "average diver" on any of the type of dives we are talking about here, most any algorithm in current use will get you out safely. However, and this is a big however, for those of us who are on the edges of the bell curve due to age, excess bioprene, low vascularity , low fitness levels etc etc, some of the algorithms may bring us much closer than we would like to be and we might be perfectly safe, right until we have a dodgy day and we get hit.

My feeling is that I personally would rather have more safety margin in my planning and if that means skewing my strategy to go shallower earlier and spend more time on the good stuff (O2 wise) then that's what I will do. The NEDU study and resultant discussions have allowed me to make an informed decision to do this, if VPM works for you by all means continue. I don't for a moment believe that either of the good Dr's on this thread have ever stated that VPM will get you bent. It just might not be the MOST efficient way of reducing DCS risk esp for those who are at higher risk.

The risk of DCS following VPM or a (sane) GF or even the old tables are pretty damn low. Im a pilot by profession though so my career won't survive even a relatively minor hit. Also, my dry job makes me think in terms of risk reduction even in very "safe" situations. Halving a 0.5% DCS risk may not sound like much until its you sucking the DAN bottle empty on the boat.

 

[RainPilot]

I think this is a very accurate summary. The one thing I would clarify is in item 1. I am not sure there was much Doppler work done in support of deep stops at the time of their popularization – at least I know of very little published data. One of the key documents often cited in support of deep stops was ”Neuman TS, Hall DA, Linaweaver PG. Gas phase separation during decompression in man: ultrasound monitoring. Undersea Biomed Res 1976;3:121-30.” This paper reported Doppler VGE measurements during decompression in a dry chamber (i.e. during the decompression stops). It is a fascinating paper because there has been relatively little work done on VGE at depth. What they found was that as a result of an unplanned “deep stop” (a short extra stop inserted 10 fsw deeper than the prescribed first stop to allow the watches to catch up after a overly-rapid first pull), there were fewer detected VGE during the decompression stops. This makes sense. However, they did not continue to measure VGE long after surfacing, so may not have detected the maximum VGE following diving – which is the VGE measurement that is correlated with the incidence of DCS (and in more recent work has generally been shown to be increased by deep stops). They did not do enough dives to find a significant difference in DCS, and most subjects got at least skin bends, as it was a very aggressive profile. A rarely cited report of the same era, “Spencer MP, Johanson DC. (Institute of Environmental Medicine and Physiology). Investigation of new principles for human decompression schedules using the Doppler ultrasonic blood detector. Technical Report. Washington DC: Office of Naval Research; 1974 Jul.”, reported a comparison of a shallow stops versus a deep stop schedule, those being a 200 fsw / 30 minute BT schedule from the Dwyer/ des Grange 1957 Standard Air decompression Table (USN Diving Manual from 1959-2008), which has a 40 fsw first stop, and the corresponding schedule from the old C&R tables (derived from Haldane’s), which has a 70 fsw first stop (and six minutes shorter total decompression time). That was a small comparison with just a couple dives on each, but the deep stops resulted in more VGE. As alluded to above, recent work – in the last decade – has indicated that compared to shallow stops, deep stops during decompression dives result in either the same or higher maximum VGE grades following diving.

Thanks for the info sir. When NAUI shifted over to RGBM all those moons ago, we had to do an instructor workshop on the new tables etc. During that discussion, one of the major reasons given for the switch-over was that the RGBM model was developed in conjunction with the WKPP, to the extent that the divers were taking computers in a pocket and after every big dive the data was downloaded and Doppler was carried out to confirm the bubble loads. This was touted as the first really scientific study of actual "recreational" divers ever and one of the few ever since Haldane et al.

Since RGBM is proprietary, is it possible that the data was never published or did my CD make it up on the day? Wouldn't be the first time that happened...

 

[The Chairman]

did my CD make it up on the day? Wouldn't be the first time that happened...

You might owe me a keyboard. That was phunni.