what do scientists say about decompression?


The foramen ovale is a small hole located in the septum, the wall between the two upper chambers of the heart. Before birth, the lungs are not used to get blood rich in oxygen. Instead, this blood comes from the mother’s placenta and is delivered through the umbilical cord. The foramen ovale makes it possible for the blood to go directly from the veins to the right atrium of the fetus’ heart to the left atrium of the heart, bypassing the lungs. Normally, the foramen ovale closes as blood pressure rises in the left side of the heart after birth, or within a few years after birth. Once it is closed, the overall blood flows to the lungs without bypass, to get oxygen before it enters the left side of the heart and gets pumped to the rest of the body.

The foramen ovale remains open, or patent (persistent foramen ovale or PFO), in about 25% of the adult population [1] (1). Most patients with a PFO do not have any symptoms. However, the condition may play a role in migraine headaches and it increases the risk of stroke, transient ischemic attack and heart attack. In divers, it has been associated with severe neurological decompression sickness, inner ear decompression sickness, and cutaneous symptoms [2]. Venous bubbles formed during and after decompression can pass through the foramen ovale shunt -so not filtered by the lung- and invade the peripheral arterial circulation. They can reach tissues supersaturated with inert gases and, as a result, there is an amplified risk of bubble emboli in these zones.


Literature overview

The various forms of DCS for which a link has been demonstrated regarding the presence of PFO have been recently reviewed [2]. Cerebral DCS seem more linked to PFO presence than spinal cord DCS [3]. Regarding brain and inner ear injuries, where supersaturation levels can be locally high, mechanisms have been proposed and discussed [2] [4]. The relationship between cutaneous forms and PFO was initially more puzzling. A recent study leads to hypothesize a cerebrally mediated mechanism [5].
The DCS risk ratio for diver with PFO compared to diver without PFO is not precisely known but several studies tend to point a two 2.5 times risk increase at least [6] and even more than five [7] [8]. Severe DCS forms are definitely more frequent when a PFO is present [9]. Additionally, the size of the PFO appears to play a major role regarding DCS risk ratio determination [7][10]. In a recent study, on the 200 divers who had an atrial defect closure (PFO and ASD) following shunt-related DCS, about 50% had an atrial defect 10mm or larger, whereas about 1% of the general population appear to have PFO diameter in this range (about 25% of the population have a PFO but with a diameter mainly between 2mm and 6mm) [10]. Nevertheless, serious DCS forms are sometimes associated with small size PFO [11].
PFO closure appears to be an efficient solution to prevent major DCS and return to unrestrictive diving [10][11], while conservative diving profiles without PFO closure seems also to lead to satisfying level of safety regarding severe DCS forms [12]. It is also associated with a decrease of asymptomatic brain lesions [13]. PFO closure surgery is characterized by a low rate of procedural complication [14].



DCS risk characterization linked to a PFO presence remains a major research topic. A rigorous statistical analysis is not yet available. A routine way of screening PFO (presence, dimension, fully, partially, intermittently opened…), with or without a DCS case history, with or without invasive means, is not feasible or even advisable.
Indeed, whereas the scientific community agrees on the fact that PFO is a DCS risk factor, a routine screening of diver is not preconized [15]. Only divers with an history of severe DCS form are eligible for a routine screening, through bubble contrast transthoracic echocardiography with provocative manoeuvres. In case of positive PFO result, PFO close using transcatheter is recommended for return to normal diving [15].
However, a conservative approach of diving, with less severe diving profiles leading to gas load limitation, appears to be an interesting alternative to limit DCS risk [16].
Finally, it must be noticed that other right-to-left shunt pathways may exist in the body, in particular at pulmonary level [17]. Currently, the DCS risk associated with this anatomical reality (with a dispersion of its importance among the population) and its role in microbubble arterialization is poorly known.



  1. Homa S, Messé SR, Rundek T, Sun YP, Franke J, Davidson K, Sievert H, Sacco RL, Di Tullio MR. Patent foramen ovale. Nat Rev Dis Primers. 2016; 2: 15086. – ABSTRACT
  2. Wilmshurst PT. The role of persistent foramen ovale and other shunts in decompression illness. Diving Hyperb Med. 2015; 45(2):98-104. – FULL TEXT
  3. Germonpré P, Dendale P, Unger P, Balestra C. Patent foramen ovale and decompression sickness in sports divers. Journal of Applied Physiology. 84(5): 1622-1626. – FULL TEXT
  4. Mitchell SJ, Doolette DJ. Pathophysiology of inner ear decompression sickness: potential role of the persistent foramen ovale. Diving Hyperb Med. 2015; 45(2): 105-110. – FULL TEXT
  5. Kemper TCPM, Rienks R, Van Ooij PJAM, Van Hulst RA. Cutis marmarota in decompression illness may cerebrally mediated: a novel hypothesis on the aetiogoloy of cutis marmarota. Diving Hyperb Med. 2015; 45(2):84-88. – FULL TEXT
  6. Bove AA. Risk of decompression sickness with patent foramen ovale. Undersea and Hyperbaric Medicine. 1998; 25(3): 175-178. – ABSTRACT
  7. Torti SR, Billinger M, Schwerzmann M, Vogel R, Zbinden R, Windecker S, Seiler C. Risk of decompression illness among 230 divers in relation to the presence and size of patent foramen ovale. Eur Heart J. 2004; 25: 1014-1020. – FULL TEXT
  8. Honek J, Sramek M, Sefc L, Januska J, Fiedler J, Horvath M, Tomak A, Novotny S, Honek T, Veselka J. High-grade patent foramen ovale is a risk factor of unprovoked decompression sickness in recreational divers. J Cardiol. 2019; 74(6): 519-523. –ABSTRACT
  9. Liou K, Wolfers D, Turner R, Bennett M, Allan R, Jepson N, Cranney G. Patent foramen ovale influences the presentation of decompression illness in SCUBA divers. Heart, Lung and Circ. 2015; 24: 26-31. – FULL TEXT
  10. Wilmshurst PT, Morrison WL, Walsh KP, Pearson MJ, Nightingale S. Comparison of the size of persistent foramen ovale and atrial septal defects in divers with shunt-related decompression illness and in the general population. Diving Hyperb Med. 2015; 45(2):89-93. – FULL TEXT
  11. Wilson C, Sayer MDJ. Cerebral arterial gas embolism in a professional diver with a persistent foramen ovale. Diving Hyperb Med. 2015; 45(2): 124-126. – FULL TEXT
  12. Koopsen R, Stella PR, Thijs KM, Rienks R. Persistent foramen ovale closure in divers with a history of decompression sickness. Neth Heart J. 2018; 26: 535-539. – FULL TEXT
  13. Billinger M, Zbinden R, Mordasini R, Windecker S, Schwerzmann M, Meier B, Seiler C. Patent foramen ovale closure in recreational divers: effect on decompression illness and ischaemic brain lesions during long-term follow-up. 2011; 97(23): 1932-1937. – ABSTRACT
  14. Pearman A, Bugeja L, Nelson M, Szantho GV, Turner M. An audit of persistent foramen ovale closure in 105 divers. Diving Hyperb Med. 2015; 45(2): 94-97. – FULL TEXT
  15. Smart D, Mitchell S, Wilmshurst P, Turner M, Banham N. Joint position statement on persistent foramen ovale (PFO) and diving. South Pacific Underwater Medicine Society (SPUMS) and the United Kingdom Sports Diving Medical Committee. Diving Hyperb Med. 2015; 45(2): 129-131. – FULL TEXT
  16. Klingmann C, Rathmann N, Hausmann D, Bruckner T, Kern R. Lower risk of decompression sickness after recommendation of conservative decompression practices in divers with and without vascular right-to-left shunt. Diving Hyperb Med. 2012; 42(3): 146-150. – FULL TEXT
  17. Lovering AT, Stickland MK, Kelso AJ, Eldridge MW. Direct demonstration of 25 and 50µm arteriovenous pathways in healthy human and baboon lungs. Am J Physiol Heart Circ Physiol. 2007; 292: H1777-H1781. – FULL TEXT

(1) In fact, there are two kinds of such holes in the heart. One is called an atrial septal defect (ASD), and the other is a patent foramen ovale (PFO). Although both are holes in the wall of tissue (septum) between the left and right upper chambers of the heart (atria), their causes are quite different. An ASD is a failure of the septal tissue to form between the atria, and as such it is considered a congenital heart defect, something that you are born with. Generally an ASD hole is larger than that of a PFO.


Ten years ago, DAN published a text that summarized perfectly the deep stops issue [1]:

“Throughout history, the key concern for divers has been how to get back to the surface without suffering decompression sickness (DCS). The British scientist J.S. Haldane combined empirical data with scientific studies to develop step-by-step decompression procedures that, along with the accumulated experience and work of many scientists, led to the development of modern decompression tables and computer algorithms.

Decompression sickness still occurs in recreational divers, but at a rate of 1 to 4 per 10,000 dives; these cases of DCS are often mild and treatable. However, the possibility of severe injury — though rare — makes divers eager to hear about any measure that might further reduce the risk. One possibility is the deep stop.

In the mid-1990s, Richard Pyle, a biomarine scientist who frequently made dives to great depths in search of fish species, noticed that sometimes he felt fatigued after dives, and at other times he felt normal. An excellent observer and trained scientist, he figured out that on dives when he had to stop during his ascent to deflate the swim bladders of his specimens, he felt much better. Soon he introduced a brief stop halfway to the surface on all his dives and formed the strong opinion that this significantly reduced his post-dive fatigue. He shared his experience with fellow divers, and the practice of deep stops became widespread among technical divers before it could be scientifically tested.

What is a deep stop? In the minds of most who practice it, the deep stop is an additional stop during ascent, introduced by divers beyond what their computer algorithm demands. However, there are now computer algorithms that claim to include deep stops, though neither these algorithms nor the practice of deep stops has been thoroughly validated.

Discussion about deep stops is not new to the scientists studying decompression safety. Since Haldane first established decompression tables, the depth of the first stop has been debated. The answers varied over time, depending on prevailing contemporary dive practices and concerns. Haldane, for example, assumed that tissues may sustain a certain level of supersaturation or critical volume of surplus gas before bubbles occur. That is why his decompression model applied a relatively quick ascent to depths he believed would drive inert gas out of the body.

Later it became clear that bubbles probably occur much earlier than Haldane assumed, and these findings led to the creation of so-called bubble models. Many dive computers on the market incorporate deeper stops than do earlier Haldanian models. Some of them are based on bubble models, while others adjust parameters of non-bubble models to achieve similar effects. However, to mimic the deep-stop practices adopted by some technical divers, some computers add stops deeper than what their models call for or give divers this option.

So the big question before divers today is: How effective are deep stops at preventing DCS, whether called for by a bubble algorithm or when used by divers independent of what their computers suggest?”

This summary remains applicable nowadays. As will be explained hereafter, the controversy on deep stop relevance is a major debate, pointing that the decompression theory is yet a “living” topic. It is a frequently discussed question in the diving field [2], with currently less firm positions: a compromise has to be found between all the pro and cons theoretical views, reflected on a practical point of view by the gradient factors adjustment issue.


Literature overview

In the early 1960’s, more than twenty years before Pyle’s observations, Brian Hills analyzed the decompression procedures used by the Pearl fishermen in the Torres Strait and introduced a new concept of decompression modeling [3]. These divers practiced their first stop deeper than commonly done at that time, with, in addition, a higher pressure drop between the last stop and the surface than previously allowed.  The zero supersaturation concept proposed by Hills assumed that bubbles can form even for very low supersaturation levels. As a consequence, the rate of the first part of the decompression has to be reduced significantly with a first stop rather deep. Even if Hills work never produced operational decompression tables, his approach inspired a generation of researchers who focused on bubble modeling and the necessity to limit the ambient pressure decrease amplitude during the ascent, in order to limit bubble generation and growth [4]. They revisited Haldane theory or at least tried to redefine the acceptable supersaturation thresholds. It was a new attractive and challenging topic.

In this context, David Yount, at the University of Hawaii, started fundamental work on decompression in the late 70’s. One of his objectives was the selection of criteria for microbubble formation, in the context of a supersaturation state. He assumed that the microbubbles forming in the body during a decompression are produced from preexisting micronuclei that are activated. Yount opened new perspectives by studying gelatin samples, compressed more or less rapidly, saturated with an inert gas and then decompressed. His experiments allowed, for the first time, the numbering of the micronuclei recruited and transformed into microbubbles under various supersaturation levels. A new decompression model called VPM (Varying Permeability Model) was finally proposed for man, using his theoretical foundations [5]. The VPM model, proposed by Eric Maiken and Erik Baker in the USA under version A then B, has been widely used by recreational technical SCUBA divers who dive rather deep with trimix and heliox mixtures: the TEK divers. It generates deep/short decompression stops compared to the more standard procedures as it limits the supersaturation levels to reduce the number of micronuclei recruited.

As a continuation of Yount’s approach and to marry several views, Bruce Wienke offered to the recreational SCUBA diving market a new algorithm RGBM widely used as a basis for several decompression computers from the 90’s.

Nevertheless, in parallel of bubble modelling motivations, Edward Thalmann works in USA for the US Navy and Albert Bühlmann works in Europe didn’t recuse Haldanian fundamentals. It is a matter of fact that they remain nowadays the basis of most of the current operational tables for both professional and recreational divers, and of most of the decompression computers.

Thalmann EL-RTA and the VVAL algorithms are of neo-Haldanian type, with exponential-linear gas exchange kinetics and various matrices of maximum permissible tissue tensions (M-value Workman approach). A slower tissue desaturation was introduced, with linear gas exchange, reflecting that bubble formation slows the inert gas removal kinetics. This constitutes a major evolution of neo-Haldanian models, pointing the fact that adaptations of the original approach are required.

In the recreational field, TEK divers, initially excited by the renewal proposed by VPM and RGBM deep stops, have finally selected Bühlmann ZH-L16 algorithm to a large extent, with the gradient factors (GF) option. They build their in-house decompression procedures with adjustment possibilities for the first stop depth (GF Low) and the total decompression time (GF High) [6]. The question of the decompression curve shape appears a central issue. This context has motivated several studies to investigate on the deep stops relevance.

In 2000’s, the French Navy tested four 50msw and 60msw air dive protocols on 12 divers with experimental ascent profiles (EAP) tested in the wet compartment of a hyperbaric chamber and with bubble monitoring to assess the benefits offered by deep stops [7][8]. The interest of deep stops was not demonstrated, with three of the EAPs showing no difference, one produced increase bubbling and one produced a joint pain DCS case compared to French Navy tables MN90.

It must be mentioned that the decompression profiles tested by the French Navy didn’t reflect the exact practice of technical divers (not fixed and clear at all) or the decompression protocol proposed by Richard Pyle [9]. This remark applies to the results observed by the US Navy through a dedicated study [10]. A 170fsw/30min experimental exposure was tested according to VVAL18 Thalmann algorithm regarding the decompression on one side, and according to a probabilistic bubble model (BVM3) producing deep stops on the other side, with the same decompression duration. DCS incidence following these two schedules were compared. DCS incidence of the deep stops was significantly higher than that of the shallow stop conventional approach. It was deduced that the slower tissue gas kinetics is impacted by a deep stop, what is easily understandable if one follows a neo-Haldanian view (1).

A more recent study tends once more to point no major advantages of a deep stop profile compared to a profile computed via Bühlmann algorithm adjusted with GF, for a 50msw/25min trimix dive: bubble production detected via 2D echocardiography was not significantly different and deep stops produced even more marked inflammatory responses [11]



It is noticeable that Pyle’s procedure, in agreement with theoretical views regarding microbubble production and growth dynamics, has never been really tested to prove its benefits compared to more conventional procedures.

As pointed by Erik Baker, there is an infinite possibility to produce deep stops even with a neo-Haldanian approach, using a Bühlmann algorithm modulated with a GF option [6]. The introduction of deep stops has to be compensated by a lengthening of the last stop close to the surface, probably to allow the egress of extra the gas load accumulated or contained in some tissues during the deep stop period.

The question of the deep stops interest becomes major if one focuses on the total decompression time required:  for a same decompression duration, is a deep stop decompression profile better than conventional shallow stop? Rare studies tend to bring a response [10][12].

DAN interest regarding deep stop benefits was initially presented in 2000 [13]. The introduction of extra deep stops appeared advantageous for what concerns bubble production. However, the lengthening of the decompression duration was significative. So, the question is the following: is it the deep stop by itself that limits the bubble production or is it the associated decompression duration increase, required according to most algorithms, that proves safer?

Additionally, if one believes that microbubble production can be minimized by opting for a deep stop approach, what is the optimal depth for the first stop, for a given diving profile?

It seems that currently, nobody can demonstrate if this optimal solution exists, as well as the total decompression duration advisable for this optimal way. Microbubbles form all the more profusely than the ambient pressure decrease is large but, in the same time, inert gas elimination is favorized by large ambient pressure decrease. So what is the best compromise? And how the inter and intra individual variabilities impact this optimal? Bubble detection -as a routine way of monitoring diver- should help to give a response to the question. A large database can be collected through the following of recreational diver activity and bubble production (large population, various diving profiles, various GF options), then analyzed.



  1. Denoble P. Deep stops. Alert Diver Online. Article 255; 2010. – FULL TEXT
  2. Powell M. Delving deeper into deep stops. Divernet. Reprinted from Diver; July 2018. – FULL TEXT
  3. Hills BA. A thermodynamic and kinetic approach to decompression sickness. Libraries Board of South Australia; 1966.
  4. Hugon J. Decompression models: review, relevance and validation capabilities. Undersea and Hyperbaric Medicine. 2014; 41(6): 531-556. – ABSTRACT
  5. Yount DE, Hoffman DC. On the use of a bubble formation model to calculate diving tables. Aviation Space Environmental Medicine. 1986; 57: 149-156. – FULL TEXT
  6. Baker EC. Clearing up the confusion about deep stops. – FULL TEXT
  7. Blatteau J-E,Hugon M, Gardette B, Sainty J-M, Galland F-M. Bubble incidence after staged decompression from 50 or 60msw: effet of adding deep stops. Aviation Space Environmental Medicine. 2005; 76(5): 490-492. – ABSTRACT
  8. Blatteau J-E,Hugon M, Gardette B. Deep stops during decompression from 50 to 100msw didn’t reduce bubble formation in man. Proceedings Decompression and the Deep Stop; Salt Lake City; 24-25 June 2008. – FULL TEXT
  9. Pyle R. The importance of deep safety stops: rethinking ascent patterns from decompression dives. SPUMS Journal. 1997; 27(2): 112-115. – FULL TEXT
  10. Doolette DJ, Gerth WA, Gault KA. Redistribution of decompression strop time from shallow to deep stops increases incidence of decompressions sickness in air decompression dives. Report NEDU TR-1106; Navy Experimental Diving Unit; 2011. – FULL TEXT
  11. Spisni E, Marabotti C, De Fazio L, Valerii MC, Cavazza E, Brambilla S, Hoxha K, L’Abbate A, Longobardi P. A comparative evaluation of two decompression procedures for technical diving using inflammatory responses: compartmental versus ratio deco. Diving and Hyperbaric Medicine. 2017; 47(1): 9-16. – FULL TEXT
  12. Schellart NAM, Brandt Corstius J-J, Germonpré P, Sterk W. Bubble formation after a 20m dive: deep-stop vs shallow-stop decompression profiles. Aviation Space Environmental Medicine. 2008; 79: 488-494. – FULL TEXT
  13. Marroni A, Cali Carleo R, Balestra C, Longobardi P, Voellm E, Pieri M, Pepoli R. Effects of the variation of ascend speed and profile on the production of circulating venous gas emboli and the incidence of DCI in compressed air diving. Phase 1. Introduction of extra deep stops in the ascent profile without changing the original ascent rates. DSL Special Project 01/2000. European Underwater and Baromedical Society, EUBS Annual Scientific Meetings; Malta; 14-17 September 2000. – FULL TEXT

(1) GF approach coupled to a Bühlmann algorithm can counter-balance this drawback by lengthen the total decompression duration.