[19] Therefore, information on stool consistency alone was also calculated (a higher number indicated a looser stool). (3) Upper respiratory symptoms: recorded using a modified Jackson system, which detailed the severity of seven items (malaise, chilliness, sneezing, sore throat, runny nose, blocked nose, and cough) each on a 0 to 3 Likert scale (the headache score was removed).[20] As there are no sufficiently

sensitive and specific clinical definitions of presence or absence of upper respiratory infections,[20, 21] the Niemen method of defining presence of upper respiratory symptoms as any score Selleckchem ABT 888 above 1 was used.[22] (4) Anxiety: recorded using the short form state-trait anxiety scale, which recorded the severity of six items on a 0 to 3 Likert scale (adapted from the usual 1–4 scale to ensure consistency with the other self-report measures).[23] Alpha coefficients for the anxiety scale in the present study ranged from 0.83 to 0.92 (above the recommended value for psychological measures of 0.70). (5) Fluid intake: determined daily by using drink bottles of known volume and bead counters to record refills. Total fluid intake was also determined from 24-hour food and fluid

diaries on day 3 (1,100 m) and day 13 (4,700 m), with food and fluid composition determined by computer software (Dietmaster; ZD1839 datasheet Lifestyles Technologies Inc, Phoenix, AZ, USA). Arterial oxygen saturation and resting heart rate: by finger tip pulse oximeter (9500, Onyx; Nonin, Plymouth, MN, USA), recorded when participants were sheltered from the wind, after wearing gloves and blinded to their results. The lowest and highest values observed over a 1-minute period were recorded and the mean calculated. To achieve the study’s first aim, for each illness, the individual symptom score and the total symptom score were calculated to provide daily expedition mean scores. Statistical Florfenicol differences between days were determined by repeated measures analysis of variance. Significant differences

were followed up by Holm–Bonferroni procedures[24] using 1,435 m as the baseline for comparison (the last day of the baseline period that exhibited normal arterial oxygen saturations). Also, for each illness, the expedition’s daily sum of symptom scores (a marker of expedition symptom burden), daily and total expedition incidence (the number of individuals achieving criteria, when available, for clinical diagnosis), and event rates (expressed per 100 person days) were calculated. Participants with missing data were removed from these analyses. To achieve the study’s second aim, longitudinal linear regression analyses were performed using generalized estimation equations.[25] The predictor variables were day of expedition, height gain, upper respiratory symptoms, stool consistency, anxiety symptoms, arterial oxygen saturation, heart rate, and fluid intake.