On trips with multiple transmeridian flights, pilots experience successive non-24 h day/night cycles with circadian and sleep disruption. One study across a 9-day sequence of transpacific flights (no in-flight sleep, 1-day layovers between flights) reported an average period in the core body temperature rhythm of 24.6 h (circadian drift). Consequently, pilots were sometimes flying through the circadian performance nadir and had to readapt to home base time at the end of the trip. The present study examined circadian drift in trip patterns with longer flights and in-flight sleep. Thirty-nine B747-400 pilots (19 captains, 20 first officers, mean age = 55.5 years) were monitored on 9- to 13-day trips with multiple return flights between East Coast USA and Japan (in 4-pilot crews) and between Japan and Hawaii (in 3-pilot crews), with 1-day layovers between each flight. Measures included total in-flight sleep (actigraphy, log books) and top of descent (TOD) measures of sleepiness (Karolinska Sleepiness Scale), fatigue (Samn-Perelli Crew Status Check) and psychomotor vigilance task (PVT) performance. Circadian rhythms of individual pilots were not monitored. To detect circadian drift, mixed-model analysis of variance examined whether for a given flight, total in-flight sleep and TOD measures varied according to when the flight occurred in the trip sequence. In addition, sleep propensity curves for pre-trip and post-trip days were examined (Chi-square periodogram analyses). Limited data suggest that total in-flight sleep of relief crew at landing may have decreased across successive East Coast USA-Japan (flights 1, 3, 5 or 7; median arrival 03:45 Eastern Daylight Time (EDT)). However, PVT response speed at TOD was faster on East Coast USA-Japan flights later in the trip. On these flights, circadian drift would result in flights later in the trip landing closer to the evening wake maintenance zone, when sleep is difficult and PVT response speeds are fastest. On Japan-East Coast USA flights (flights 2, 4, 6 or 8; median arrival time 14:52 EDT), PVT response speeds were slower on flight 8 than on flight 2. Circadian drift would move these arrivals progressively earlier in the SCN pacemaker cycle, where PVT response speeds are slower. Across the five post-trip days, 12 pilots (Group A) immediately resumed their pre-trip sleep pattern of a single nocturnal sleep episode; 9 pilots (Group B) had a daytime nap on most days that moved progressively earlier until it merged with nocturnal sleep and 17 pilots (Group C) had nocturnal sleep and intermittent naps. Chi-square periodogram analyses of the sleep propensity curves for each group across baseline and post-trip days suggest full adaptation to EDT from post-trip day 1 (dominant period = 24 h). However, in Groups B and C, the patterns of split sleep post-trip compared to pre-trip suggest that this may be misleading. We conclude that the trends in total in-flight sleep and significant changes in PVT performance speed at TOD provide preliminary evidence for circadian drift, as do persistent patterns of split sleep post-trip. However, new measures to track circadian rhythms in individual pilots are needed to confirm these findings.
