Peripheral Perfusion Versus Lactate Targeted Fluid Resuscitation in Septic Shock

GENERAL OBJECTIVE To demonstrate that a CRT-targeted FR is associated with less positive fluid balances, organ dysfunctions, and similar improvement of markers of tissue hypoperfusion and hypoxia, when compared to a lactate- targeted FR.

SPECIFIC OBJECTIVES

1. To determine if normalization of CRT is associated with less fluid resuscitation and positive 24-h fluid balances as compared to lactate-targeted resuscitation.
2. To determine if normalization of CRT is associated with less organ dysfunctions as compared to lactate-targeted resuscitation.
3. To determine if CRT-targeted resuscitation compared to lactate-targeted resuscitation leads to similar improvement in markers of tissue hypoperfusion or hypoxia at the end of fluid resuscitation
4. To determine if normality of all the selected hypoperfusion variables (central venous oxygen saturation (ScvO2), central venous to arterial carbon dioxide pressure difference (P(cv-a)CO2), and CRT) at different time-points in septic shock patients can predict the absence of tissue hypoperfusion and hypoxia.

For this Research Project, several variables of tissue hypoperfusion that can only be assessed by special perfusion-monitoring techniques were included.

Sublingual microcirculatory assessment: Sublingual microcirculatory alterations have been well described in septic shock patients. Functional capillary density and microvascular blood flow are decreased, while heterogeneity is markedly enhanced. These alterations have been shown to be prognostic, with a rapid improvement in survivors but a progressive decline in nonsurvivors. A microcirculatory flow index (MFI) ≤ 2.5 and a proportion of perfused vessels (PPV) < 80% will be considered as categorical of microcirculatory hypoperfusion Liver blood flow: Plasma disappearance rate (PDR) of indocyanine green (ICG), with a non-invasive transcutaneous assessment of ICG clearance. Normal range is 18% to 25% per minute with a value < 15%/min categorically abnormal.

Muscle tissue O2 saturation: Muscle tissue O2 saturation (StO2). A decrease in this variable to <70% suggests profound tissue hypoperfusion.

Assessment of tissue hypoxia: There are two clinically calculable variables that have been proposed as closely representing tissue hypoxia: the venous- arterial CO2 (Cv- aCO2) to arterial-venous O2 content (Da-vO2) difference ratio (Cv- aCO2/Da-vO2) and the lactate/pyruvate (L/P) ratio . Both ratios are an expression of anaerobic metabolism at the cellular level and thus can be linked to hypoxia. For this research project, these variables were considered to ascertain the presence and resolution of tissue hypoxia.

• Cv-aCO2/Da-vO2 ratio: This ratio might be useful as a surrogate of the respiratory quotient. A ratio ≥ 1.4 could identify anaerobic CO2 generation. A high Cv-aCO2/Da-vO2 ratio in the setting of hyperlactatemia may favor anaerobic metabolism as the possible source of lactate, while a normal Cv-aCO2/Da-vO2 ratio suggesting that lactate accumulation is due to non-hypoperfusion- related causes.
• L/P ratio: In anaerobic conditions, pyruvate is transformed to lactate and thus the L/P ratio increases to ≥ 18. The L/P ratio is considered to be one of the most reliable indexes of hypoxia in critically ill patients.

METHODOLOGY A randomized controlled study of parallel groups was designed: Group A with a CRT-targeted fluid resuscitation strategy and group B with a lactate-targeted one.

This prospective study will be performed at the Hospital of the Pontifical Catholic University and at the Public Hospital of the South of Santiago, Chile.

The study was approved by the Institutional Review Board of both centers. A signed informed consent will be asked to the next of kin of all eligible patients, and confirmed by the patients when feasible.

The study intervention period will be of 6 hours I. Randomization A randomization sequence with an allocation of 1:1 will be generated by a computer program. Study-group assignment will be performed by means of randomized permuted blocks of eight. Allocation concealment will be maintained by means of central randomization.

Fluid challenges (500 ml of Ringer's lactate administered in 30 min intervals) are repeated until the perfusion target is normalized, or fluid responsiveness becomes negative, or a safety limit of an increase in central venous pressure (CVP) ≥ 5 mmHg after a fluid bolus is reached. The perfusion target in group A is a normal CRT ≤3 sec. The perfusion target in group B is an arterial lactate ≤ 2 mmol/l or a decrease >20% every 2h.

CRT will be assessed every 30 minutes and lactate every 2 hours during the 6h study intervention period, after which the treatment is liberalized for attending physicians.

If fluid resuscitation is stopped because the perfusion target is normalized, the time is registered, and the patient subjected to the specific research-related assessment protocol and followed until hospital discharge.

II. General management algorithm Besides sepsis source aggressive management, all patients will be treated according to our local algorithm aimed at macrohemodynamic stabilization and improvement of hypoperfusion. The algorithm has been described elsewhere. Co-interventions will be registered and considered in statistical analysis.

III. Specific research-related assessments For the purposes of this research protocol, several variables will be periodically measured or calculated as follows: at baseline, at 2h, at 6h and at 24h.

1. Macrohemodynamic variables such as mean arterial pressure,M, heart rate, norepinephrine dose, CVP, dynamic predictors.
2. Continuous cardiac output (CO) monitoring: this will be performed with non-invasive pulse-contour CO assessment acquired with this project (PiCCO device).
3. Metabolic-perfusion variables such as arterial lactate, ScvO2, and P(cva)CO2.
4. Peripheral perfusion assessment: CRT and mottling score
5. Sublingual microcirculation: It will be assessed with the side dark field (SDF) device. At each assessment, at least five 10-20 sec video images will be recorded. The analysis will be performed by eye following recent recommendations . From image analysis the following variables will be calculated (i) proportion of perfused vessels; (ii) microcirculatory flow index (MFI); All these indexes will be calculated separately for small (<20 microm diameter) and large vessels (> 20 microm diameter) .
6. Liver blood flow: An ICG finger clip well be fixed in every patient and then connected to a liver function monitor (LiMON; Pulsion Medical Systems, Munich, Germany). A dose of 0.25 mg/kg of ICG (LiMon Pulsion Medical Systems, Germany) will be injected through a central venous catheter.
7. Near infrared spectroscopy (NIRS): StO2 will be measured by a tissue spectrometer (InSpectra Model 325; Hutchinson Tc, Mn, USA). A NIRS probe will be placed on the skin of the thenar eminence
8. Ccv-aCO2/Da-vO2 ratio: This ratio will be calculated after taking arterial and central venous blood gases with the Douglas formula.
9. L/P ratio: This ratio will be assessed at 0, 6 and 24h during the study period, at baseline and when fluid resuscitation is stopped. Arterial blood samples for pyruvate will be taken and processed in our laboratory before 3h according to the method described by De Backer et al , including immediate deproteinization of the sample and analysis by enzymatic fluorometric-assay (Sigma-Aldrich, USA)
10. Sequential organ failure assessment (SOFA) at baseline, 24, 48 and at 72h . Finally, all patients will be followed until hospital discharge, and all data including demographic aspects, sepsis sources and management, inflammatory biomarkers, and severity scores and major outcomes will be registered.

STATISTICAL ANALYSIS Sample size calculation: the simple size calculation was based in some small clinical studies. In a recent study, a resuscitation strategy guided by peripheral perfusion offered important benefits in terms of fluid administration at 6 hours (4227 ± 1081 ml vs 6069 ± 1715 ml) and organ dysfunction, when compared with standard fluid therapy. In consequence, a 1600 ml difference in the mean 24-h total fluid administration between the experimental CRT group and the conventional lactate group was considered to be the critical threshold for hypothesis testing. If there is truly no difference between the standard and experimental treatment, then 46 patients are required (23 patients per arm) to be 90% sure that the lower limit of a two-sided confidence interval will be above the limit of -1600 mL at an alpha level of 0.05. Statistical analysis Intention to treat analysis will be performed in order to minimize bias.

Analyses will be performed as follows: Specific Objective #1: between study groups, the amount of fluids administered at the end of fluid resuscitation, and at 6, 24, 48 and 72h will be compared exclusively at each time-point with t-test, and comprehensively using ANOVA. Total 24-h fluid balance will be compared using t-test. Specific Objective #2, the status of peripheral perfusion normalization will be assessed in a categorical way (normal/abnormal) with chi-square at specific time-points (at the end of fluid resuscitation, 6, and 24h), and with means of SOFA score at each time-point up to 72h using t-test or Wilcoxon rank sum test if appropriate. A total comparison using ANOVA will be also performed. In the same line, the relative change of SOFA among time-points (with baseline value as the reference) and between study groups will be assessed with analysis of proportions, using the binomial distribution. Specific Objective #3: comparison of improvement on markers of tissue hypoperfusion and hypoxia between study groups will be mainly assessed in a categorical way using chi-square at different time-points, since those markers are qualitatively attributed a dichotomic normal/abnormal status. For some of them (ICG, StO2) as continuous variables, t-test or Wilcoxon rank sum test will be used at specific time points (at the end of fluid resuscitation, 6, 24h), and ANOVA for a global analysis. Specific Objective #4: For this objective, pooled data from all the studied patients will be used. Pearson correlations between selected hypoperfusion variables (ScvO2, P(cv-a)CO2, and CRT) with markers of tissue hypoperfusion/hypoxia will be explored at different time-points.

All statistical calculations will be performed using Stata Statistical Software, Release 14 (College Station, TX). A probability value (p-value) of less than 0.05 will be considered