The Effect of Restrictive Fluid Management on Cardiac Function and Glycocalyx Degradation

This study is a substudy to the larger CLASSIC-trial that aims to investigate restrictive and standard fluid therapy in treating septic shock. Adult patients with septic shock according to the Sepsis-3 criteria who have received at least 1 L of IV fluid in the 24 hours before screening will be screened. Patients who have had septic shock for more than 12 hours at the time of screening, who have life-threatening bleeding, or acute burn injury >10% of the body surface area, who are pregnant and those in whom consent cannot be obtained will be excluded. Blood samples will be drawn at T0 (during the first hour after enrolment), T1 (the first morning after inclusion) and T2 (the second morning) and T3 (at ICU discharge (within 24 hours before discharge)). The samples will be analyzed for high-sensitivity troponin T (hsTnT), pro-BNP, proAdrenomedulin (MR-proADM), Co-Peptin (AVP), endothelin-1 (ET-1), neuregulin-1 (NRG-1), growth differentiation factor-15 (GDF-15), metalloproteinases (MMPs), hyaluronan, syndecan-1, heparan sulfate, IL-6, TNFR and Ang-2.

Systolic and diastolic function parameters for the left and right heart will be collected at study enrolment (within 24 hours from inclusion) and at 2-3 and day 7-10 or at discharge using transthoracic or transesophageal echocardiography.

Baseline clinical and demographic data will be analyzed using chi-square or Fisher's exact test for categorical data, and the Wilcoxon rank-sum test for continuous data. The baseline and first follow-up measure of the concentration of a biomarker will be modelled using a mixed effect linear model with a person specific random effect and an interaction between time of follow-up measure (interval since first measurement) and treatment group, the stratification variables for the randomization: hematological or metastatic cancer (Y/N) and trial site as fixed effects. A mixed model will be used to analyze the pattern of biomarkers over time. Differences between intervention groups will be tested by the interaction of time and group in the same mixed model analyses. Additional covariates will be added to adjust for treating center, illness severity (SMS-score), cumulative fluid balance and comorbidities.

The investigators have based sample size calculations on both primary outcomes, hsTnT and hyaluronan concentrations, and estimated predicted differences and standard deviations found in previous comparable studies. The Jakobsen-Lange will be used to adjust for two outcomes where αi for each outcome is αi=0.05/((n+1)/2) = 0,033 which is an adjustment halfway between no adjustment and full Bonferroni adjustment with n being the number of co-primary outcomes and secondary outcomes respectively. For sample size calculations a power of 80 % will be used. Using a standard deviation of 40 ng/l of hsTnT this reveals a sample size of n=120 to detect a group difference of 22 ng/l. To account for drop-out and loss-to-follow-up an extra 10% will be added, 132 patients will be included equally distributed between the two groups. Using a standard deviation of 29 ng/ml of hyaluronan concentration reveals a sample size of n=225 to detect a group difference of 11,5 ng/ml. To account for drop-out and loss-of-follow-up an extra 10% will be added, 248 patients will be included equally distributed between the two groups.

Inclusion criteria

• Patients with septic shock according to the Sepsis-3 criteria who have received at least 1 L of IV fluid in the 24 hours before screening

Exclusion criteria

• Patients who have had septic shock for more than 12 hours at the time of screening
• Patients who have life-threatening bleeding
• Patients with acute burn injury >10% of the body surface area
• Pregnant patients
• Patients in whom consent cannot be obtained