Airline management representatives would not permit flight attendants to either to carry and activate the VN air samplers, or to complete pre- and post-flight health surveys. Instead, some members of the research team tested the feasibility of carrying the samplers onboard on commercial flights on which they were booked. This occurred during 80 flights (1.5 – 5 hour sampling duration, 0.4 – 0.9 l/min flow rate), and these samplers were returned to the lab for analysis. None of the researchers reported either adverse air quality or symptoms on the limited number of sampled flights.

The first two rounds of air sampling on a total of nine commercial flights are difficult to reliably interpret, in part because of evidence that some samplers had not been cleaned properly prior to distribution, and were contaminated independent of the aircraft. In the third round of sampling, the samplers were thoroughly cleaned prior to distribution, and at least one TCP isomer was detected in 14 of the 38 inflight samples, with none of the field or lab blanks above the detection limit for any TCP isomer. (Blanks are intended as “controls” to test that contaminants detected on the filter came from the cabin air and not from some other source). The positive samples were collected on the following aircraft types: A319, A320, B737-800, B757, CRJ-100, and ERJ145.

In addition to the three rounds of sampling collected above, duplicate samples were collected on 18 flights to check for inter-laboratory agreement. Here, again, these data are difficult to interpret, in part because the analytical equipment at one of the three labs was programmed to a detection limit which was considerably higher than the anticipated range of sampling data concentrations. The agreement between the other two labs was inconsistent, but both labs found detectable TCP isomers in multiple samples under what were apparently normal operating conditions. (A systematic error at one of the labs resulted in the measurement of ooo-TCP in every sample including controls. This was traced to interference from a separate experiment that had been run with the equipment).

The VN sampler did identify nanogram (ng) range levels of tricresyl phosphates under normal operating conditions on commercial flights. Also, the combination of TCP isomers identified in the chromatographic fingerprint is consistent with the TCP blends added to aviation engine oils, so is a valid marker of oil fumes on aircraft. The inter-laboratory results suggest that the reliability of the data reported for specific isomers in the nanogram range was not always strong, but the reliability should be higher when recovering the higher amounts of TCPs anticipated during “actual” or visible fume events. Overall, the data collection and inter-laboratory comparison exercise provided valuable insight into procedural safeguards necessary to ensure the validity of sampling data. Additional air sampling should be conducted on a larger sample of flights to better characterize the prevalence and range of exposure, including on incident flights where fumes are noticeable and/or accompanied by visible smoke/haze.