2023 marks the 70th anniversary of the first published report that I know of about synthetic oil fumes in the bleed air being a problem for crewmembers. If anyone is counting, that’s 25,394 days. This 1953 report describes B52 pilots reporting symptoms inflight coincident with the presence of smoke and odor sourced to the air supply. They noted that the contamination was the worst immediately after the air conditioning system was turned “on.” And returned “intermittently during engine power changes.” A lot of things have happened since 1953. QEII has been coronated. There was an uprising in East Berlin against Soviet occupation. And Americans were driving the Hudson Hornet, watching “From Here to Eternity.” There have been so many tremendous technological leaps forward since 1953 – the integrated circuit all the way up to ChatGPT and the ISS, with an honorable mention for the B787. But why aren’t bleed filters and sensors and more non-bleed systems also on this list? Fox News, July 27, 2018: „Spirit Airlines flight diverted due to hazmat (hazardous material) situation after passengers sickened from 'dirty socks' odor“ (https://perma.cc/FH4A-GGMX). Why are we still here? In the aviation industry, there is, what seems to be a long-standing practice of promoting uncertainty on the question of whether breathing engine oil fumes can either compromise flight safety or cause ill health. The oft-repeated claim is that “there is no consensus on this cabin air quality issue,” which – on its face – is true. A recent review paper by Hayes et al. made that claim on their review of a subset of the literature which found that 80% of the papers which acknowledged crew experience or perspective concluded that chemicals in cabin air did pose an occupational health risk, but only 5% of the papers which acknowledged manufacturers or airlines drew that conclusion. The vast majority of those who claim either “no problem” or “not sure that there’s a problem” are either employed by or consult for the industry. And even if that’s done with the best of intentions, the optics are not good. In contrast, crewmembers are clear that oil fumes are a problem. Now, some of you may be thinking, “This is a complex technical issue – it takes time to solve.” And others may be thinking, “Is it even a problem?” Some of you may even be undecided. But for 70 years now, crewmembers - whose only commercial interest is that they depend on their airlines to be financially viable so that they can work – have been saying that the air isn’t reliably clean, it’s not clean enough. And they’ve also been saying that, when there are oil fumes in the air supply, some people get really sick. And sometimes, flight safety is compromised. Given that it’s been 70 years now, it is a good time to delve more deeply into the issue of delay, starting with a definition. Delay is “a tactic of slowing down a decision-making process in order to maintain the status quo.” You may know about the field of “motivation theory” where academics discuss what motivates behavior. I came across a researcher at the University of Calgary who coined an equation to describe motivation which I have simplified here as: DCT = ES · VC ÷ ATD DCT (Desire to Complete Task) ES (Expectation of Success) VC (Value that you place on Completion) ATD (Acceptable Time Delay) Generally - and not universally - we have seen limited desire from regulators, manufacturers, and airlines to complete the task of preventing exposure to oil fumes. I say that because fume events still happen. All but one aircraft type uses engine bleed air. And bleed filters and sensors barely exist. That is the status quo which is being maintained. And in the context of this equation, part of that “low desire for getting the job done” is about the numerator: Some filtration companies may want to develop a marketable product but may question if they can succeed, especially without an aircraft manufacturer as a reliable partner and without guaranteed airline customers. Similarly, aircraft manufacturers probably don’t see much “value of completion” – why spend money to make changes that aren’t required, which would make your product more expensive and, therefore, less competitive? The biggest part of the minimal progress to complete the task, though, is the denominator – the ATD. We are at 25,394 days and counting. Primary school math taught us that when the denominator in a fraction is big then the value of the fraction is small. The task at hand here is to “prevent exposure to oil fumes on aircraft”. It is my view that there have been some powerful and effective tactics to slow down any decisions. These are important to acknowledge so that we can all focus, instead, on getting the job done. I will start by saying that, in fairness, delay tactics were not the initial response to oil fumes. The language in that 1953 report in my first slide shows that, initially, industry swiftly treated fumes in the air supply as an engineering problem that required an engineering solution. Boeing acknowledged that the fumes and smoke were from engine oil, especially during power setting changes. They said the bleed temperature influenced oil breakdown. And said that “without doubt, the materials in question are numerous and complex.” So, they injected oil into a bleed air simulator and tested different bleed air filtration options – activated charcoal, particulate filters, catalysts – much like we talk about today. They predicted having an “excellent opportunity of success.” But by 1957, Esso proposed a different response. Esso had run some tests with guinea pigs breathing fumes from heated oil (temps from 400-900F). Some of the GP died, others had degenerative changes in the brain, lung, liver, and kidneys, esp. at higher temperatures. Still, Esso said they weren’t sure which chemicals in the mixture caused the toxic effects. And, even then, they said not to worry because the concentration of each individual chemical wasn’t high enough to be a problem. That is, they applied what sounded logical to counter their own observations of guinea pigs getting sick and dying to justify the status quo. Esso also cited some UK MOD human exposure data in which pilots reported irritated eyes and headache but could still “carry out normal functions”. Esso concluded that oil fumes “should be eliminated” but were “principally a nuisance.” Note: Esso Turbo oil 15 was first oil used in bleed systems per Jean Szyudar, French expert on aviation engine oil history. VC went down. The ATD went up. Delay #1: Claim that the symptoms are minor, and pilots can still fly the plane. Therefore, it’s safe. But the problem was that the problem of oil fumes didn’t go away. Even with smoking permitted during commercial flights which masked many odors, cabin crew at AA and UA were reporting routine headaches on the DC10 associated with a distinctive and strange smell. (Three PW JT9D engines on DC-10-30 vs. PW JT8D engines on DC-9). Those airlines, to their credit, hired a consultant who concluded that the most likely source of the smell and the headaches was pyrolyzed oil in the bleed air stream – from leaking oil seals, and so on. Presumably, this information was shared with the manufacturer of the DC10, Douglas Aircraft Corporation. 1973 would not have been the first time that Douglas had heard about oil fumes. In 1966, Douglas wrote an internal memo about whether engine bleed air was suitable for ventilation on civil aircraft. They said that various fatal crashes had been attributed (rightly or wrongly) to oil fumes. But they also said that, to certify aircraft, they’d simply need to show that pilots could smell the fumes before the fumes were dangerous. And if pilots can smell the fumes, then they can isolate the fumes. Problem solved. Four months before the first DC-10 flight by AA, Douglas, Esso, and Humble Oil had met with FAA to discuss oil fumes on the DC10. Those three companies told the FAA that odor is detectable long before CO will degrade crew performance. And, they said, the levels of CO and aldehydes were below workplace exposure limits, so it was fine. Again, VC went down and ATD went up. Delay #2: Pilots will smell fumes before it gets bad. And the FAA certified the aircraft. So, it must be safe. Reports kept coming. During the 1990s, pilots in Australia reported ill health and also impairment while flying the Bae146 aircraft. And they spoke loudly enough that a Senate committee conducted a two-year enquiry, including eight in-person hearings. And the Senate report concluded that… -- Pilots have been and can be impaired while flying this aircraft -- compromising flight safety. And they said that, even though this aircraft seems worse than others, the problem applies to all commercial aircraft because, at the time, all aircraft used unfiltered bleed air. As a result, they recommended a long list of exposure control measures. And it is interesting to read some of those submissions to the Senate Committee. The Senior VP of BaE said that, with the weight of human evidence and suffering … The director of the Civil Aviation Safety Authority said that oil fumes are a feature of the basic design of systems that use bleed air for ventilation. Ultimately, all that came of that two-year Senate enquiry and its recommendations was that – on a precautionary basis only – British Aerospace made mandatory what had been a recommended practice of inspecting and cleaning the ducting, and ECS for oil. Aviation regulators, eventually, also mandated those inspection and cleaning procedures for this aircraft type. But the design of that bleed air system - or any bleed air system – didn’t change. Delay #3: Accept that oil fumes are “a feature of the basic design” of bleed air systems. Meanwhile, in the Northern hemisphere, also in the year 2000, the US Congress told an NRC committee to conduct a one-year independent study, both to identify contaminants in the aircraft air and to develop recommendations for ways to reduce such contaminants, including – they specifically said - a non-bleed design (Sec. 725 of PL 106-181). In its report published two years later, the NRC committee told the FAA that it needed to “rigorously demonstrate the adequacy of current and proposed regulations related to Cabin Air Quality (CAQ) and revise if inadequate.” And to their credit, the FAA agreed. They said that they may not have “kept pace” with public expectation and concern about Air Quality – that their current rules don’t “afford explicit protection”. The FAA said that … At that same time, though, Boeing had drafted a research proposal for what they called a “Multiple Factors” (MF) study – noting that crew and pax were reporting sx onboard but how much of that is air quality-related and how much of it is from other factors like stress and an uncomfortable seat?” Boeing enlisted interest and some funding (ultimately $225K) from an engineering association called ASHRAE. Importantly, the “Multiple Factors” study proposal that Boeing drafted was not intended to address what they called “upset conditions” – or fume events. They said so, explicitly, in their approved work statement. But the following year, when Congress told the FAA to fund researchers to collect swab samples of ventilation duct residues and recirc filters after fume events, specifically in response to the NRC report – one of two research groups included HSPH and they were directed to collect their onboard measurements as part of the MF study. In effect the FAA called for a “two-fer” – two for the price of one. Boeing, ASHRAE, and the FAA would pool their resources, collecting sensor data and survey data on normal flights – not expected to have a fume event - to contribute to the MF study and (at the same time) ostensibly to respond to Congress. (FAA Reauthorization Bill of 2003, Section 815 - became Public Law 108-176). In 2004, two years after promising an air quality rulemaking that would address fume events, the FAA Federal Air Surgeon testified before Congress that the agency had intended to initiate rulemaking on cabin air quality, but they’d defer that promised action because of the Boeing Multiple Factors study – the study which had nothing to do with fume events. FAA told Congress that they would circle back to regulatory action on CAQ in 2006 or 2007 after they had reviewed “substantial data” from the MF study. Boeing also testified that day about what had become known as the “ASHRAE/FAA Multiple Factors” study – one of those factors was airborne contaminants when there is no fume event. Plus, noise, humidity, fatigue, and others. 8 and 14 years later, respectively, the FAA and ASHRAE published the sampling data from 83 flights, none of which included a reported fume event, designed by Boeing and funded by ASHRAE and American taxpayers. There were findings on air quality and there were findings on other factors. As an example, an “other factors” finding was that: “flights in the southern hemisphere were associated with lower odds of reporting pain in the neck and shoulders”. And cramped seating is uncomfortable. On the air quality front, the overall conclusion was that “the air quality and environmental conditions in the passenger cabin of commercial airplanes are comparable to or better than conditions reported for offices, schools and residences, with a few exceptions.” Which is, perhaps, why the FAA didn’t circle back. There has been no subsequent regulatory action to address fume events. The VC went down again. And the ATD went up. To be clear – the MF study would have been okay if it was just privately funded. And a public-private partnership might even be okay if the research goals had been the same. It would even be okay if the partnership had only been to collect ozone and pesticide data on aircraft, because Congress had asked for that, too. The problem was that the focus of the research that the law called for and the focus of the MF study were very different – but the way it was worked out, public funds were used for research which cancelled a promised FAA rulemaking to help to prevent exposure to oil fumes. Delay #4a: Delay regulation with research and design the research to not address/define the problem (at all). There are other versions of long and expensive publicly funded research projects that sidestep the repeated requests from crew members to prevent fume events. For example, in Europe, around 2013, the European Commission offered 2 million euros to fund a five-part study on fume events and health. An academic and industry team collected measurements on 69 flights, not making any effort to target fume events, and no fume events occurred on the sampled flights. Still, the final report assures the reader seven times that Occupational Exposure Limits (OELs) weren’t exceeded during the 69 flights. I think that’s supposed to be comforting. But it’s not relevant to fume events. And that’s a lot of money to not learn anything new. The team also tried to provoke two “artificial fume events” during ground ops but acknowledged that they didn’t measure much, and nobody reported symptoms. Finally, the team generated oil and hydraulic fumes and exposed lung and brain cells for either 30 minutes or 24 hours. Those results WERE interesting – in some of the trials, the oil fumes influenced neuronal activity and HF fumes were toxic to the lung cells. But, since then, nothing seems to have come of those findings. The authors of the FACTS report did call on EASA to review the way it certifies aircraft air supply systems. That’s because air quality certification is a one-shot deal prior to the aircraft entering service, but this team questioned whether that approach ensures sufficiently high-quality ventilation air over the 30-40 years that the aircraft is in service?” To my knowledge, EASA has not yet acted on that recommendation. Finally, the authors noted that “several exposure control measures are possible” but that’s “up to each individual player in the field.” And I agree with that statement in FACTS report – it IS up to each individual player in the field to control exposure to fumes. It shouldn’t be, but nobody seems to be in charge. If you think back to that equation, the VC for industry is low because there are no meaningful regulations to meet. Delay #4b: Delay regulation with research that inches the field forward a little (but not much). For decades now, the response to crew and flight safety reports during and after inhaling oil fumes onboard has typically been to measure oil fume constituents in cabin air during non-fume event conditions. I think of this as determining whether the concentration of smoke during a house fire poses a hazard by measuring the concentration of smoke in the kitchen when the house is not on fire. They’re not the same. Still, even the “normal flight” measurement data do show something. They show that, routinely, there can be low levels of TCPs and other oil and hydraulic fluid components in the cabin air. But instead of using that information as evidence of the need to filter and monitor bleed air to notify pilots on the location of contamination in the system so that they can isolate in real time and then direct maintenance to the source of the problem upon landing, these reports just compare “non-event” data to OELs. And as you might expect, the comparison makes the cabin air look pretty good. Some examples to express how long this approach has been used. You can read… “well below the generally accepted tolerance limits”, results “did not reveal a health hazard = potentially harmful dose is not possible via inhalation at levels at or even at higher levels”, “below the established CAA, Federal Aviation Administration (FAA) and contractual limits” etc. The central claim of almost every such study is that the measured contaminant levels during normal conditions are below OELs or other guideline values, and thus, exposure to fumes during a fume event can’t explain the reported symptoms. Just like Esso first said in 1957 when they exposed GP to oil fumes and had to explain why the oil fumes weren’t the reason that the GP got so sick. Again, the VC goes down, ATD goes up. Delay #5a: To determine if fume events are a problem, measure airborne contaminants when there isn’t a fume event and compare the data to limits that don’t apply. [Note that OELs are also irrelevant to the acceptability of supply air quality during a fume event.] Other proposed explanations for crew symptoms include panic attacks, hyperventilation, nerve sensitivity, and the nocebo effect. On balance, though, it is hard to conceive of thousands of crew members inventing an illness that has zero benefit for them. If you get sick from breathing oil fumes, there are financial implications – lost income, denied Worker Compensation (WC) claims, and big medical bills. Sometimes, crews are transported to the Emergency Room directly after a flight and the bill isn’t covered by WC. There are also professional implications. Cabin crew can get disciplinary points when they are too ill to continue a trip. Too many points and they’re fired. Pilots sometimes lose their license to fly. In the US, where health insurance is typically tied to your job, if you can’t work, you lose your health insurance. And then, if you need medical care, you lose your savings and your home. Crew unions don’t make money on fume events, either. Also, crewmembers are an organized, take-charge work group, but they’d have to really go above and beyond to coordinate their stories with other crew members, globally, over a period of 70 years. If they hated their job to start with and, upon pretending to be sick, were given excellent medical care and awarded large sums of money without condition, then maybe a conspiracy theory would be reasonable. But, under the circumstances, actual illness does seem like a more likely explanation. Delay #5b: Rule out the chemicals and then propose psychiatric causes. One seemingly dull and administrative change that would help is to establish a fume event reporting rule. And ALSO to tweak the existing mx reporting rules. The current reporting rules for fume events only require airlines to document the mechanical defect or failure that generated fumes and generallly are only reportable when the wheels of the aircraft are up. That’s a problem because the mechanical source of a fume event may not be found, and some fume events aren’t even caused by a mechanical defect, and most fume events are reported to start on the ground (sourced to the APU) when the wheels are down. So, a lot of fume events don’t have to be reported. Also, there is no reporting system for crew members or mechanics who inhale fumes and report sx. Congress has told the FAA to create such a reporting system not once but twice – in 2003 and 2012. But still, somehow, twenty years on, we still don’t have one. Delay #6: Just do nothing. Something that has really grown in the past 10 years is recognition of the need to train crewmembers to recognize and respond to fume events. Because if you think about it, if there is smoke sourced to fire onboard, then some of the procedures – like donning oxygen masks – are the same for fume events. But some of the procedures are different. With smoke sourced to fire, for example, increasing the ventilation flow to dilute the contaminants should help. But if the supply air is the source of the oil fumes, or if you just think that a dirty sock smell means that someone has taken their shoes off, then increasing the air flow will make the fumes worse because you are also increasing the supply of fumes. In 2015, ICAO published a circular which outlines what pilots, flight attendants, mechanics, and high-level management need to know to recognize and respond to oil and hydraulic fluid in the supply air specifically because of the flight safety concern when crews are impaired by fumes inflight. ICAO also said that it’s important for airline workers to be trained to document fume events in a standardized reporting system. Note that this circular is available for purchase on the ICAO website. On the industry side, Airlines For America (A4A) has led on the worker training front. In mid-2018, they invited all stakeholders – airlines, crew and mechanic unions, regulators, and manufacturers – to join an A4A task group and draft a spec to define practices to recognize, mitigate, and report cabin air quality events (all kinds, including oil fumes). And it was good because there was a range of perspectives in the room, and we all worked together. Unfortunately, the spec was published about two weeks before the world shut down because of covid so it didn’t get the attention it otherwise would have – although it continues to be available online at no charge. Also, in 2018, in line with the increased recognition of the importance of training, Congress approved “Sec. 326(a), Public Law 115-254, 2018”. It calls on the FAA to post educational materials online for crews and mechanic workers about how to recognize and respond to fume events. The language is not great – the FAA still need not require airline worker training, but it could have been a start. FAA announced a website in March 2021. Unfortunately, the materials posted are not directed to educating cabin crew directly, with the possible exception of the ATA Spec. The ICAO circular is listed under “trade organizations,” but it’s misattributed to the IATA. And there’s no information on what’s in the ICAO report or how to purchase it. And none of the relevant educational materials developed by some crew associations is included. The webpage is not responsive to the 2018 law. Delay #7: Don’t train crews to recognize and respond to fume events. If people don’t know what’s happening during an event, the VC stays small and the ATD stays big. AFA represents Cabin Crew at 18 airlines; of those, one (Norse Air) has a bleed-free fleet but, of the other 17, two have SOME Flight Attendant training on fume events (Alaska and Spirit) and two are considering it for 2023. The rest have nothing. There are many more examples and I could go on, but I know that the conference director is tracking my time. So, to close, I’d like to shift gears and focus on what needs to change. I’ve listed what I think are the nine key items: Bleed free designs or Less toxic oils and hydraulic fluids Effective bleed air filtration/cleaning technology Real-time detection systems Mandatory crew education and training Fume event pilot checklists Standardized fume event reporting systems Preventive maintenance protocols Disease recognition All of these are described in CEN Technical Report 17904 published last year. Which you can buy online. ** So, how will these NINE things happen, you ask? Well, to circle back to the equation, the only way to raise VC and ES for everyone is for aviation regulators around the world to require bleed systems to be equipped with real-time sensors and filters. Because for crews – VC remains high, clearly. But for industry, it’s still low. And if bleed air systems were properly regulated, the ATD would need to be small. It’s not zero, the industry can’t solve this problem overnight from where they are today. But we need a time limit to bound it, and we need the time limit to actually get enforced. Otherwise, we will be waiting from here to eternity. In closing, the industry has not been motivated to regulate itself. And, in a way, why would it? We need regulatory agencies to regulate. We need to stop funding research that muses over the question “How much oil fumes are okay to breathe?” We need to collectively focus on the task at hand. Until regulatory change happens, we urge airlines to be a more discriminating customers with Boeing, Airbus, Embraer, and others by telling them that you want and would buy bleed-free systems, bleed air filtration, useful sensors, and less toxic oils. We also urge airlines to improve their maintenance and reporting practices, and to train crews appropriately. In effect, to truly put safety first. I hope you look for and call out the delay tactics listed in this slide going forward. And after three days of presentations, I hope each of you return to your day job to actively work to create the market and the environment for change. For cigarettes, it took more than 100 years of research, discussion, and debate to fully recognize the health risks of smoking. The aviation industry can do better.
This is a publication from the International Aircraft Cabin Air Conference 2023 (Online, 27-29 June 2023)
