Methods/design
The trial is a pilot, efficacy, single-center RCT comparing initial plasma with BC resuscitation in surgical patients with septic shock. The trial will investigate modulation of SHINE as a strategy by which outcomes in patients with shock may be improved. The trial will follow the CONSORT (Consolidated Standards of Reporting Trials) guidelines15 and has been registered on ClinicalTrials.gov (NCT03366220).
Setting
The study will take place in the Emergency Department (ED) and 23-bed Shock Trauma Intensive Care Unit (STICU) at Memorial Hermann Hospital-Texas Medical Center, a level I trauma center located in Houston, Texas.
Study population
The target population includes critically ill, traumatically injured or surgical patients who have septic shock. Patients who meet the inclusion criteria of ≥18 years old and have a Sepsis Screening Score ≥416 with a suspected source of infection (figure 2) will be eligible for enrollment. Enrolled patients with hypotension with mean arterial pressure <65 mm Hg, and signs of hypoperfusion such as lactic acid >2.2 mmol/L, altered mental status or decreased urine output (<0.5 mL/kg in the past hour) will be randomized to receive either crystalloid or plasma resuscitation. Exclusion criteria are listed in figure 3.
Figure 2The Sepsis Screening Score. Score of <4 out of 16 has a 96% negative predictive value for sepsis.16
Figure 3Patient exclusion criteria.
Screening and enrollment
Patient screening and enrollment will take place in the STICU and ED (figure 4). In the STICU, on-call residents and fellows will communicate with research personnel when a patient has suspected sepsis. Once notified, the research team will evaluate potential patients for enrollment eligibility. Enrolled patients in the STICU will be monitored for up to 72 hours (in time for cultures to result) by the research team for development of septic shock. On meeting these criteria for septic shock, patients will be randomized and immediately started on, or switched to (if patient is already receiving fluids) therapy fluid (either plasma or BC). In the ED, research assistants are available 16 hours/day to screen patients who meet the enrollment criteria.
Figure 4Patient randomization flow sheet from Shock Trauma Intensive Care Unit (STICU) and Emergency Department (ED). MAP, mean arterial pressure; UOP, urine output.
Crystalloid administration may already be under way prior to patient randomization. Patients who have received the entire 30 mL/kg crystalloid dose prior to enrollment are not eligible to participate in this study. Written informed consent will be obtained from subjects or a legally authorized representative. Once consent has been obtained, and patient meets randomization criteria, prior fluids will be exchanged for therapy fluid (either plasma or BC).
Randomization, allocation concealment and blinding
Randomization will occur once the patient meets the criteria. One-to-one allocation will be used for the randomization to either the intervention arm (plasma) or the control arm (BC). Randomization will be performed using a computer-generated random sequence placed in opaque, consecutively numbered, sealed envelopes kept in the locked research office. A research team member is available 24 hours/day to consent and randomize eligible patients. Patients will be stratified by enrollment location (STICU vs. ED) in five blocks of 4 and one block of 6 for a total of 26 patients. The healthcare provider will not be able to be blinded to the study intervention. The outcome assessors will be blinded when feasible, and the laboratory technicians and statisticians will be blinded.
Intervention
Type and screen will be performed to provide donor-matched plasma for subjects randomized to intervention group. Initial resuscitation with plasma will be 10 mL/kg (700 mL in a typical 70 kg adult). Traditional doses of plasma, when used to correct coagulopathy, range from 10 mL/kg to 15 mL/kg.17 Plasma will be administered at a rate of 2 mL to 3 mL/kg/hour (140 mL to 210 mL/hour in a typical 70 kg adult). After the initial dose of plasma has been given, subsequent fluid resuscitation (crystalloid or colloid) will be given at the discretion of the treating clinician. For patients who proceed to the operating room during fluid resuscitation, the anesthesiology team will continue fluid administration per study protocol.
Control
Usual care using BC (Iso-Lyte) only will follow Surviving Sepsis Campaign guidelines. Controls will receive 30 mL/kg (2100 mL in a typical 70 kg adult) of crystalloids within the first 3 hours.5 Crystalloid administration may be terminated before the entire dose has been administered if patients show clinical improvement, or if the treating clinician has concerns about circulatory overload. After the initial dose of crystalloids has been given, subsequent fluid resuscitation (crystalloid or colloid) will be given at the discretion of the treating clinician.
Endpoints of resuscitation: Fluid resuscitation in both the intervention and the control arms will be titrated to serial reassessments of the patient’s volume status at the discretion of the treating clinician. Serial laboratory examinations will be performed per standard of care in the ICU. Additional fluid support after the initial plasma or crystalloid bolus will be given at the discretion of the treating clinician. A research physician will be at bedside to follow patient resuscitation. Plasma administration may be terminated before the entire dose is administered if patients show clinical improvement, or if the treating clinician is concerned for circulatory overload.
Outcomes
The primary outcome is a reduction in serum biomarkers, soluble-thrombomodulin, which is associated with glycocalyx breakdown and endothelial injury at 6 hours after initiation of fluid therapy in the plasma group. Biomarkers will be drawn at 0 and 2 hours at study fluid administration completion, 6, 12 and 24 hours to evaluate their trend in response to plasma versus crystalloid resuscitation. Lactic acid will be drawn per standard of care to guide resuscitation. Standard labs for patients with critical illness will be obtained per ICU protocol.
Secondary outcomes include additional volume of fluid required for resuscitation after initial bolus of study fluid within the first 24 hours of resuscitation, time on vasopressors, time until lactate normalization, ventilator days, ICU-free days and hospital length of stay. Organ dysfunction will be measured including acute lung injury and acute renal failure. Standardized definitions will be used as outlined in the method of operation used in previously conducted trials by our group.18 Potential harms of plasma administration will be assessed. Risk/safety evaluation will be performed at half recruitment by the Data Safety and Monitoring Board, which consists of a general surgeon, pulmonary critical care physician and a statistician. The trial will be terminated early if a series of adverse events attributable to plasma transfusions were to occur.
Sample size calculation and statistical analysis plan
We will enroll 26 patients to obtain unbiased estimates of treatment effect. We chose 26 patients based on the annual average number of patients with septic shock treated in the STICU assuming a 75% enrollment and randomization rate, and the goal of completing enrollment in 1 year. We will calculate estimates of treatment effect and 95% CIs for all measures. We will perform both a per-protocol and an intention-to-treat analysis, and employ both frequentist and Bayesian methods. For the primary outcome and all continuous outcomes, we will use mixed models that include group and time period as covariates with a random subject effect. We will report group differences and 95% CIs from these models. For binary outcomes, a log binomial will be used to estimate relative risks and 95% CIs. HRs and 95% CIs will be reported for time-to-event comparisons. For the Bayesian analysis, neutral conservative priors will be used to estimate the probability of both benefits and harms, as defined by our primary and secondary outcomes.