Statistical Analysis

Data were analysed using GraphPad Prism version 6.07 for Windows (GraphPad Software, San Diego, CA, United States). Sample size calculations were performed using a proprietary determined sample-size calculator using GraphPad StatMate version 2.00 for Windows. Data inspection and the D’Agostino-Pearson Test was undertaken to assess normality of all continuous data, which were presented as mean ± standard deviation (SD) and median [interquartile range] for highly skewed data, respectively.

Categorical variables were compared using Chi-squared tests. Comparison of 10 and 300 s RMSSD values were compared using the Wilcoxon matched-pairs signed-ranks test. Comparisons of continuous data over the four altitudes were performed using a one wave ANOVA and Tukey post-tests and the Kruskal–Wallis Test (non-parametric ANOVA) with the Dunn post-test as appropriate. Correlations were performed using Pearson and Spearman rank correlation (±95% confidence interval, CI) as appropriate.

The association between BP⁺ 10 s and the CMH 300 s RMSSD HRV measures were examined using correlations. Their level of agreement was investigated using a modified Bland–Altman relative-difference plot method as previously described (19–21). This was performed as the 10 and 300 s RMSSD values and their differences were highly skewed and not normally distributed. The differences between the 300 and 10 s RMSSD measures (y-axis) were compared with the average values from the comparative readings (x-axis). The 95% limits of agreement were estimated using the 2.5th and 97.5th percentile of the differences (300 s − 10 s) in RMSSD values and the average bias estimated as the median of the differences. A good agreement between the comparative RMSSD was defined as ≥95% of readings falling within these limits (21). The differences between the 300 and 10 s were further examined by altitude. Binary Logistic regression [odds ratio (OR) and 95% CI] was used to examine the relationship between AMS (yes/no) and HRV.

A two-tailed p-value < 0.05 was considered statistically significant for all comparisons.