Fig. 1From: Untangling glycaemia and mortality in critical careCohort selection from original 371 patients who treated according to the Specialised Relative Insulin Nutrition Tables (SPRINT) protocol. The first comparison compares survivors and non-survivors from Cohort 1, using as much data as possible and excluding patients with a very short intensive care unit (ICU) stay. The second comparison uses Cohort 2 to assess the impact of competing risk due to patient dropoutBack to article page