On 17 November 2017, an Airbus A320 (EC-HQJ) being operated by Vueling on a daytime scheduled international passenger flight from Geneva to Barcelona stopped its climb at FL290 after both pilots had begun to feel nauseous and when the situation did not improve, a PAN was declared advising crew incapacitation. A diversion to Marseilles was then flown without further developments but with both pilots on oxygen and they were both hospitalised for tests. There were no similar instances amongst the 154 occupants in the passenger cabin.
An Investigation was carried out by the French Civil Aviation Accident Investigation Agency, the BEA, based on downloaded FDR and CVR data, recorded ATC data and crew statements.
It was established that with the Captain acting as PF, the aircraft had taxied to runway 05 at Geneva behind a Netjets Cessna Citation Excel. It had then been stationary behind this aircraft at the holding point with the wind calm for almost four minutes until it was cleared to line up and take off. The distance between the two aircraft whilst stationary was approximately 70 metres. Once the Citation had departed, the A320 was then similarly cleared to line up and take off.
Shortly after the landing gear had been retracted, the flight crew were recorded discussing “the thick exhaust gases emitted by the Cessna Citation” which had disappeared when the Citation had lined up. The Captain remarked that he “felt almost nauseous because of them” and the First Officer confirmed that it was a very strong smell. Soon after this, whilst climbing through FL100, the Captain asked the Geneva APP controller what type of Cessna Citation was ahead of them but was informed that there was no traffic ahead of them at that time. The Captain replied that there had been a Cessna Citation in front of them on departure and that he intended to file a report due to fumes on the flight deck. The First Officer then told the Captain that he also felt “unwell and intoxicated” and ATC advised that the aircraft ahead at departure, which had taken a different route thereafter, had been a Cessna Citation Excel.
Half a minute later, the Captain was recorded telling the First Officer that he did not feel well and suggested to increase the air intake. The flight was then transferred to Marseilles ACC and was cleared to climb to FL290. The SCCM was called to the flight deck to check if there was any evidence of a similar problem in the passenger cabin but he advised that there was not. The First Officer then said that the exhaust fumes from the aircraft which had been in front of them whilst taxiing out had made him feel physically ill and the Captain then asked the SCCM to both leave the flight deck open and to remain in close proximity and “keep an eye on both him and the First Officer”. Less than five minutes later, the Captain told the First Officer that he was starting to feel nauseous and on the latter’s suggestion, the Captain donned his oxygen mask. The First Officer then briefly left the flight deck to use the toilet facility and on his return remarked that he was now feeling a little better but had previously felt very nauseous. A discussion then took place about a suitable diversion should the situation worsen and Marseilles was selected. Seven minute after the Captain had donned his oxygen mask, the First Officer did likewise and they both kept them on for the remainder of the flight.
Just over 20 minutes after takeoff, and having levelled at FL290, the Captain declared a PAN and requested a diversion to Marseilles, specifying that it was because of crew incapacitation which was approved with a subsequent request to squawk 7700. With runway 31 in use but the wind calm and ‘CAVOK’ (cloud and visibility OK) being reported, the crew asked if they could land on runway 13. This was also arranged and radar vectors to the ILS for that runway were subsequently provided. Whilst intercepting the ILS LOC, the Captain and First Officer were recorded discussing their respective states and whilst both said that they felt a little better, they decided to keep their oxygen masks on. A stable approach was flown with the AP remaining engaged until 195 feet agl and the landing was uneventful.
The emergency services were in attendance on the apron and both pilots reported that they had “continued to feel nauseous” despite opening their side windows. They were both then taken directly to hospital, from where, after ‘normal’ results from a blood test and a short period of monitoring, they were discharged two hours later. The blood test showed no evidence of carbon monoxide poisoning, the samples were not kept and no other toxicological analysis was carried out. It was noted that the general medical care available had not included any provision for detailed medical investigations into the cause of the condition presented or retention of biological samples taken because this was, in line with normal procedures, only done if it was judged to have a bearing on the immediate treatment required. This situation meant that data potentially useful to the Investigation was not obtained. Both pilots reported that “nausea and dizziness alternating with periods of respite nevertheless continued for several days”.
Possible Origins of the Contamination
The Investigation was able to provide a potential explanation as to why the effects experienced were confined to the occupants of the flight deck by comparing the supply of air bled from the engines and/or APU to the passenger cabin with that to the flight deck.
It was noted that the A320 air supply system was of a design generally typical of almost all those installed in pressurised aircraft in current service. In more detail:
- Filtered bleed air is mixed with previously-filtered re-circulated cabin air in an approximate ratio of 60:40 with the selected air temperature having very little effect and the mix being identical for the flight deck and cabin air supplies.
- The origin of the bleed air in each case differs in that 60% of the air fed to the flight deck comes from the left engine air conditioning pack and 40% is re-circulated cabin air whereas the 60% of the bleed air fed to the cabin comes from both the left and right engine air conditioning packs with a greater proportion coming from the right engine pack.
- The airflow per seat is slightly higher in the flight deck than in the cabin both on the ground and in the air and flight deck air is renewed approximately every minute whereas passenger cabin air is renewed approximately every two to three minutes.
The principal possible source of air contamination other than the Citation exhaust fumes (or other potential external causes of which there were none) was acknowledged as the engine bleed air from which the cabin air originates, which was explored to the extent possible.
In any aircraft turbine engine, seals are installed to confine the lubricants of the rotating parts and it is generally acknowledged that wear or damage to these seals may allow small quantities of lubricants to enter the bleed air in the compression system after which part of this pyrolysed lubricant can then pass through the pressurisation system and enter the cabin. The primary lubricant, engine oil, includes around 3% Tri Cresyl Phosphate (TCP), a small proportion of which may, when pyrolysed, become Tri ortho-Cresyl Phosphate (ToCP), ortho-isomers of the TCP and neurotoxic organophosphorus compounds. Although there is no scientific consensus on the matter, it was noted that some studies have suggested ToCPs could be the cause of some cabin air contamination events. Although such contamination may be accompanied by a smell of “wet socks”, this is often imperceptible and in the absence in many cases of this or any other perceptible smell, such an occurrence can only be identified by detecting associated symptoms.
In an attempt to seek more enduring evidence which would indicate the possibility of TCP or ToCP contamination, it was noted ToCP fixes itself to certain plastics and may leave traces in air filters. Samples were therefore taken from the aircraft avionics and air conditioning filters, the flight deck sun visors and flight deck seat table but specialist laboratory testing was negative for all the samples. It was noted that hair is the only biological material capable of fixing substances of interest at the time of contamination since it retains them for several weeks or months. Samples of the potentially contaminated hair strands were taken from all four aircrew after sufficient subsequent growth had occurred (seven weeks after the event) to be able to compare the part of a hair strand which would have been exposed to contamination with a part of the same strand which had since emerged from the scalp. Appropriate analytical toxicology techniques were used but no substances of interest were found in quantities which would correspond to a one-off significant exposure such as during the flight concerned and although contaminants identified included TCP, the concentration was only at levels typical of normal environmental pollution with no differences between the quantities observed in the “before” and “after” hair sections. The possibility of a TCP or ToCP contamination source was therefore excluded, although it was accepted that this did not rule out the possibility that the pilots were contaminated by some other substance inducing effects similar to those encountered and which rapidly disappeared.
The likelihood of exhaust gas contamination from the Citation efflux was therefore further examined. CVR data confirmed that there was a direct connection between the fumes perceived by the pilots and the presence of the Cessna Citation ahead of them on the ground. It was noted that the centrelines of the Citation engines, which are mounted on the rear fuselage, and that of the A320 engines is at an approximate height of two metres above ground level. The relative location of the two aircraft and the prevailing calm wind conditions were undoubtedly conducive to the formation of a concentration of exhaust gases in front of the A320 engine intakes and were this effect to have been greater in the case of the left A320 engine, a greater primary source concentration in the flight deck may have resulted.
Burnt jet fuel exhaust is made up of CO2, H2O, N2, O2 and SO2 and in certain combustion conditions may also contain nitrogen oxides (NOx), CO, unburnt hydrocarbons and volatile organic compounds (VOC). It was observed that NOx has a distinctive acrid smell and SO2 a distinctive “stinging” smell and noted that the crew statements characterised the fumes they had experienced as both “acrid” and “acidic”. However, it was recognised that these may not necessarily be the chemical compounds which caused their symptoms. In particular, CO was noted to have effects consistent with those experienced, can result in incapacitation and is without smell. The negative result of the blood tests carried out at the hospital could be explained by the time which had elapsed before the sample was taken and by both pilots’ intake of mask-sourced oxygen which could have eliminated or considerably reduced the rate of carbon monoxide in the blood.
It was also noted that Netjets had assisted in establishing that there was no evidence of any airworthiness issues such as oil leak which could have resulted in the exhaust fumes emitted by the Citation engines being in any way abnormal.
Cabin Air Quality Studies
The Investigation reviewed a number of studies relevant to the continuing prevalence of occurrences in which flight crew performance is reported to be negatively affected by contaminated cabin air:
- In May 2014, the BFU Germany published a review of over 600 events between 2006 and 2013 involving smoke, fume or odours. This found that although 180 of the events “had an impact on the health of those affected”, none of these had significantly impacted flight safety. The review included Safety Recommendations on improving the identification of and response to cabin air contaminants which may be hazardous to health, the standardisation of notification procedures, related improvements in aircraft type certification processes and research by a suitably qualified institute into the possible link between fume events and the long term health impacts.
- Statistics compiled by the UK CAA in respect of almost 1,700 fire, smoke and fumes occurrences reported during a two year period commencing in late 2014 which found that 5% may have been related to engine bleed air and in 10% no aircraft related cause was found and the focus was on an external origin. No meaningful conclusion could be reached in 36% of the reports.
- The European Union Aviation Safety Agency (EASA) has taken the view that, in general, the evidence available shows that cabin air quality is similar or better than that in normal indoor environments and that there is no scientific evidence of any link between exposure to cabin air contaminants and subsequently reported health symptoms. A study it commissioned in 2014 defined a method of measuring cabin air contamination and used this to conduct measurements during the 12 month period beginning in July 2015 on 69 flights which included 8 different aircraft/engine combinations. All except eight of these flights were on aircraft with conventional bleed air systems, the exceptions were those on Boeing 787 aircraft. No occupational exposure limits were exceeded on any of these flights. A second study investigated the toxicity of engine oil contamination via the aircraft bleed air system and concluded that whilst neuroactive products are present, “their concentration in the presence of an intact lung barrier is so low that it did not represent a major concern for neuronal function”. The final reports of both of these studies were published in March 2017.
The Conclusions of the Investigation included the following:
Despite the wide range of actions undertaken, the Investigation was not able to factually identify what caused the flight crew’s symptoms and physical discomfort. The hypothesis of them having inhaled an excessive quantity of carbon monoxide, contained in exhaust gases emanating from the Cessna Citation which had been in front of the A320, is consistent with the information collected and can explain the symptoms observed (dizziness and nausea). Nitrogen oxide and sulphur oxide compounds present in exhaust gases may also have contributed to the acrid and irritating odours smelt while taxiing. However, it cannot be excluded that the crew were intoxicated by another substance which either quickly disappeared or which was not specifically searched for in the samples taken from the aircraft as to date, not identified, even in the most recent studies.
Toxic substances were searched for in the hair samples using the most effective, innovative techniques to date. The BEA believes that the use of these techniques on matrices such as saliva, blood or urine, sampled as quickly as possible after the symptoms, and in particular as soon as possible after landing, would increase the chances of detecting a wider range of potentially toxic substances.
A safety investigation into flight crew incapacitation caused by the inhalation of toxic products could provide information by trying to establish the link between the presence of chemical compounds and the symptoms reported by crew members. This action is in addition to other areas of study and research which may be carried out by other organisations.
The occurrences for which safety investigations are opened must be targeted in order to optimise the use of substantial investigation resources. The criteria for undertaking the investigation should include the salience of the symptoms and the presumed accessibility of factual data.
Investigatory Lessons Learned
The comprehensiveness of the Investigation led to a number of useful observations which should be borne in mind if a productive investigation into the origins of a similar cabin air fumes-related event is being considered:
- As it is necessary to collect biological samples from crew members at an early stage, the safety investigation authority involved cannot act in isolation. Implementation of prior local agreements between aircraft operators and airport authorities and in particular dedicated airport emergency medical services where these exist would enable etiological medical examinations on crew members to be performed without delay.
- To be effective, this type of response would require preparation so that agreements with aircraft operator and air crew representatives could be put in place so that the conditions in which samples would be taken are acceptable. It may also be useful if potential synergies with appropriate health care facilities in the vicinity of an airport to which affected crew could be directed were identified.
- Occurrences which do not become the subject of an independent safety investigation could be the subject of an identical response process if aircraft operators or safety regulators deemed this appropriate.
The Final Report was published simultaneously in both English translation and in the definitive French language version on 3 November 2020. No Safety Recommendations were made.