An Analysis of Rebreather Deaths

Closed-circuit rebreathers were introduced to sport divers in 1998, but it wasn't long before they were associated with a number of deaths. Given the small number of rebreather units out there, compared to open-circuit scuba, the number of rebreather-related deaths seemed out of proportion. It raised the question that there may be some factor intrinsic to rebreather use that increases the risk of death. To answer that question, Andrew W. Fock, head of the department of intensive care and hyperbaric medicine at the Alfred Hospital in Melbourne, Australia, decided to look at rebreather fatalities recorded between 1998 and 2010 by Deep Life, a British manufacturer of rebreather gear.

Since 2007, Deep Life founder Alex Deas has tried to document all known civilian rebreather deaths in an Internet database. The information appeared to be derived largely from the internet forum Rebreather World. Reports on the site's "accident forum" were not independently vetted, but they had details of the victims and analysis of the events. Then in 2008, Divers Alert Network (DAN) held a technical diving forum and invited prominent members of the dive industry to discuss this database and its consequences. The scrutiny revealed significant inaccuracies in several cases, so they agreed to review the database and cases reported, correcting errors and adding information on the remaining cases.

Using this updated Deep Life database, holding discussions with training agencies and rebreather manufacturers, and then getting additional rebreather death data from the British Sub-Aqua Club (BSAC) and the DAN Asia-Pacific database, Fock sought to answer some key questions: 1) what's the rate of rebreather diver deaths compared to normal sport diving; 2) what are the major causes of rebreather deaths; and 3) what changes should be made to rebreather training and/or design to minimize future deaths?

Rebreather Numbers and Death Rates

Between 1998 and 2010, 181 rebreather deaths were recorded in the Deep Life database. There was a peak of 24 deaths in 2005 (prior to that, deaths averaged eight per year and afterwards, deaths averaged 20 annually.) Between 1995 and 2011, the three major training agencies conducted 18,000 entry-level rebreather certifications, and 500 each of intermediate and advanced certifications annually during that same timeframe. Thus, Fock estimates that in 2010, there were approximately 14,000 active rebreather divers worldwide, making an estimated 30 dives per year. At an annual death rate of 20 divers per year, this equates to a death rate of 4 per 100,000 dives per year -- 10 times that of non-technical sport diving.

Of the 181 rebreather fatalities recorded by Deep Life, 80 were attributed to equipment problems, 43 to diving-related problems, and 57 had insufficient data to form any conclusion. In the 27 deaths recorded in the BSAC study, 14 cases were associated with either equipment failure or the unit being turned on incorrectly; in only five cases was the cause of death thought to be unrelated to the type of dive gear used. Looking at all the cases, Fock believes two-thirds of them are associated with high-risk behaviors.

The Most Common Risk Factors

Fock writes that it's hard to estimate the number of active rebreather divers because manufacturers are unwilling to divulge the number of units sold, due to liability risks. And connecting risk of death to a rebreather type or brand is also hard because the number of units sold doesn't necessarily represent the number of units now in active use. Nonetheless, Fock stands by his death rate of 4 per 100,000 dives. This makes rebreather diving five times more dangerous than hang gliding and 10 times more than horse riding (but eight times less dangerous than base jumping).

The BSAC study was of importance because it showed that British rebreather divers were four times more likely to be in a fatal accident than open-circuit divers -- they only represented four percent of all dives but 14 percent of all fatalities. Also, 38 percent of deaths were associated with diving to depths greater than 130 feet, regardless of what dive gear was used. Diving below that level represented 11 percent of total dives in that study, equating to a three-fold increase in risk of death associated with just depth alone. With the assumption that rebreathers are used for deep, mixed-gas diving, this raises the issue as to what extent the breathing apparatus itself is responsible for increased risk, and to what extent it is a function of a deeper, more dangerous environment.

Two types of cases appeared in the Deep Life database most frequently: divers attempting very deep dives with limited experience, and divers continuing to dive despite rebreather alarms indicating problems with their units. Despite more than a decade of warnings, the dangers of overconfidence don't seem to have been taken to heart by many new rebreather divers. Furthermore, there have been a number of near misses reported on Rebreather World forums that seem to arise from misinformation via the Internet. Those issues continue to be a challenge for rebreather safety.

While much of the increased mortality associated with rebreather use is related to high-risk behavior and the risks of diving at depth, the complexity of rebreathers means they are more prone to failure than standard scuba gear. The risk of purely mechanical failures results in a theoretical increase of failure risk to 23 times that of a standard open-circuit tank setup. But redundancy in some sub-systems can reduce the risk of failure, particularly in key areas like electronics. For example, a rebreather that has two computers with twin batteries has a lower failure risk than that of a simpler rebreather with its single O2 display. And the ability to "plug in" off-board gas via totally independent mechanism, which some rebreathers have, reduces the risk of mission-critical failure by threefold.

Human Errors

There is little data on the actual mechanical failure rates of both open-circuit and rebreather dives, but failures are commonly reported on Internet forums. Analyzing human factors in rebreather failures, Fock estimates that more than half of them were attributed to poor training or poor pre-dive checks. The experienced open-circuit diver who takes up rebreather diving is at particular risk of overestimating his ability. With open-circuit diving, there is usually only one correct response to failure. But the complexity of rebreather diving with its interaction of physics, physiology and equipment means there may be many possible responses that allow the diver to keep breathing, but not all of them will result in a successful outcome.

Fock uses this case as an example: A diver entered the water with his rebreather turned off. He had pre-breathed the unit before entering, but not for enough time for the oxygen partial pressure (PO2) to fall to a critical level. His descent resulted in an increase in PO2, despite the consumption of oxygen from the loop. At 45 feet, he became aware that the electronics were not turned on. His options were: 1) bail out by switching to open-circuit scuba; 2) ascend to 20 feet and flush the rebreather with oxygen to get a breathing mix that's non-hypoxic on the surface; or 3) turn on the electronics (not recommended because the unit would try to calibrate the oxygen cells underwater, but it would be possible if the correct sequence was followed).

While the PO2 in the rebreather's loop was still quite breathable, an understanding of physics and phiysiology would have told him that to ascend without the addition of oxygen would result in a rapid fall of PO2 in the breathing loop. He was an experienced open-circuit diver, so his first reaction was to return to the surface to correct the problem. But he became unconscious from hypoxia just below the surface and drowned. The entire event occurred in less than 150 seconds after he started the dive. There was nothing wrong with his rebreather. Rather, he failed to make a pre-dive check to verify that the electronics were turned on, and did not take enough pre-breathe time. This type of problem can occur when a diver has completed standard checks but then delays the dive while making some adjustments, e.g., re-siting the shot line. He may respond by turning off the unit in a misguided attempt to save battery life, then fail to turn it on again because he was distracted with "getting on with the dive." The scenario above was eminently salvageable without the need to go "off the loop," but the failure to understand the consequences of the various options resulted in a tragic outcome.

What Can Be Done

The use of basic checklists and "good design" can eliminate the chance of human error wherever possible. In Fock's opinion, good design should make the execution of action and the system's response visible to the diver, use constraints to lock out possible causes of error, and avoid multimodal systems.

Training should hammer home the basic skills so that they become "hard wired" into the diver's brain, thereby allowing clear thinking in times of stress while making critical decisions. Fock believes that one way to do this is not to offer decompression diving in the initial rebreather certification course, which would then mean only a limited failure response would be required, similar to open-circuit diving (like opencircuit bailout as the only option). Once the actual diving skill set and basic rebreather management is well ingrained, then more complex teaching of physics and physiology can be introduced in conjunction with discussions on alternative bailout options and decompression diving.

"Analysis of Recreational Closed-Circuit Rebreather Deaths 1998-2010," by Andrew W. Fock, Diving and Hyperbaric Medicine, June 2013, pgs 78-85.