Statistical Analysis
The primary end point was 28-day mortality in the groups Se+and Se-. A sample size calculation was carried out using these parameters: alpha-level 0.05, statistical power level 0.8, anticipated effect size (f2) of 0.06, corresponding to the small-to-medium effect size. This resulted to a suggested sample size of 150 subjects.
As secondary endpoints, plasma Se levels were evaluated over time in respective groups to monitor the effect of Se supplementation. Whole blood GPx activity, biochemical and inflammatory markers were assessed at selected timepoints for the effect of treatment over time, and to compare the effect between groups.
Two separate post hoc subgroup analyses were made, evaluating mortality in dichotomized subgroups of patients with higher vs. lower initial APACHE II scores (above and below the median of the group), and in dichotomized subgroups with higher vs. lower SOFA scores.
Intention-to-treat analysis approach was exercised for all parameters except in patients that were excluded based on the a priori set criteria.
To test the differences in quantitative parameters between Se- and Se+groups, the non-parametric Mann-Whitney test was used. To test the quality, indicators were used depending on χ-square test and to find statistically significant correlations, non-parametric Spearman’s correlation coefficient was used. The level of significance was set at 0.05. The statistical software package SPSS 13.0 for Windows was used.
Laboratory tests
Plasma Se, whole blood GPx activity, C-reactive protein (CRP), procalcitonin (PCT), prealbumin, albumin and cholesterol levels were measured at baseline (BL, on the day of admission) before the first dose of Na-selenite, and on the days 1, 3, 5, 7, 10 and 14. An Acute Physiology and Chronic Health Evaluation (APACHE) II and SOFA scores, (grading the underlying disabilities and the severity of current illness, respectively) were determined at the same timepoints and at 28days, along with SIRS severity (SIRS and sepsis, severe sepsis, septic shock).
All routine biochemical parameters as albumin, prealbumin, CRP, cholesterol, were determined with standard clinical-chemistry methods recommended by the International Federation of Clinical Chemistry (IFCC). PCT were measured by enzyme-linked immunosorbent assays (MRP8/14 Bűhlmann and Vidas Brahms, respectively). Plasma Se was determined by atomic absorption spectrometry with electrothermic atomisation (AAS). Whole blood GPx activity was measured using the UV method (RANDOX). All blood samples were collected at 6:00 am. Clinical scoring was performed at 10:00 am.
Fig. 1 The participant flowchart diagram
Exclusion criteria:
· expected hospital stay less than 5 days (e.g. acute intoxication, transient postoperative care)
· persistent vegetative state after hypoxic brain injury or traumatic brain injury
· pregnancy
· participation in another clinical trial
· refusal to participate or inability to get an informed consent
· patient with limited care (e.g. do-not-resuscitate advance directive on file)
Fig. 4a Correlation between plasma Se levels and SOFA score at admission: P = 0.001; P (Se+) = 0.004,
P (Se-) = 0.042. SOFA, sequential organ failure assessment.
Fig. 4b Correlation between plasma Se levels and SOFA score on day 7: P=0.748; P (Se+) = 0.826;
P (Se-) = 0.512. SOFA, sequential organ failure assessment.
Fig. 4c Correlation between plasma Se levels and SOFA score on day 14: P = 0.831; P (Se+) = 0.507;
P (Se-) = 0.643. SOFA, sequential organ failure assessment.