Statistics

All statistical analyses were carried out with SPSS version 20. The Shapiro-Wilk test was used to test for normality of the sample distribution, and the Levene’s test was applied to verify homogeneity of variances to ensure the validity of analysis of variance calculations.

To examine the effect of ketamine on systolic/diastolic blood pressure and heart rate, 3 separate MANOVAs were carried out to compare the ketamine with the placebo group (hypothesis I). We used the measurement points 0 minute to 40 minutes because of the constant measurement conditions (supine, resting participants). Follow-up ANOVAs for single timepoints within this observation period were corrected for multiple comparisons with Holm’s sequential Bonferroni method. The between-group effects for measurement points 60 minutes and 120 minutes were tested using separate 1-way ANOVAs. Due to missing measurements at TP0, TP5 was chosen to provide a baseline for the heart rate analysis. Too few heart rate measurements were documented at TP60 and TP120, resulting in necessary exclusion of these timepoints.

To elucidate whether blood pressure/heart rate changed significantly from baseline, we additionally computed mixed-design ANOVAs with the 2 factors time (0 minutes, 40 minutes) and group (ketamine, placebo), applying the Greenhouse-Geisser correction followed by simple effect analysis (Bonferroni-corrected). For heart rate analysis, we used timepoints 5 and 40 minutes. This method allowed us not only to reveal a difference between the 2 treatment groups but also to show an increase/decrease within the groups compared with the initial values before treatment.

To identify factors influencing the physiological parameters, we defined 2 variables to describe the blood pressure and heart rate curves yielding a single value for each participant. The maximal change in the curve during the observation period (ΔMAX) and the time to achieve this maximal change (TΔMAX). Both have unique clinical predications and have been described in the literature as providing a reliable statistical analysis of serial measurements (Matthews et al., 1990) (supplementary Figure 1).

Pearson’s correlation was calculated to study the relationship between the baseline (TP0) of systolic/diastolic blood pressure and heart rate and the ΔMAX and TΔMAX (hypothesis II).

To investigate the effect of the NET polymorphism, a mixed-design ANOVA was carried out (group factor: ketamine, placebo; genetic factor: [AA], T-carrier) for the systolic blood pressure (ΔMAX_Sys and TΔMAX_Sys) followed by a simple effect analysis (Bonferroni-corrected). To assess potential gender related differences, we performed a mixed-design ANOVA (group factor: ketamine, placebo; gender factor: male, female; time factor: TP 0–40 minutes) for the diastolic blood pressure. Additionally, we examined ΔMAX_Dia and TΔMAX_Dia for the gender factor separately for the ketamine and placebo groups in a simple effect analysis (hypotheses III, IV).

Finally, we performed a multiple linear regression analysis with the independent factors NET polymorphism and baseline systolic blood pressure at timepoint 0 (TP0_Sys) on the dependent factor TΔMAX_Sys.