Materials and Methods
This open-label randomized placebo-controlled parallel-group trial was conducted between October 2010 and March 2012. Recruitment was initially from the three mixed medical-surgical adult ICUs within Imperial College Healthcare NHS Trust (based at Charing Cross, Hammersmith, and St Mary's Hospitals). In addition, patients were recruited at Guy's and St Thomas' NHS Foundation Trust adult ICU from December 2011 to February 2012. Independent research ethics committee approval was obtained (10/H0604/35). In view of the emergency nature of the trial, a waiver of initial consent was granted. Patients could be enrolled into the trial without upfront consent, but then, consent was obtained from the patient or a personal or professional legal representative as soon as practically possible. In cases where a legal representative gave consent, retrospective consent was sought once the patient regained capacity.
Patients
The inclusion criteria were adult patients (≥ 16 yr) who had sepsis (2/4 systemic inflammatory response criteria due to known or suspected infection) and who required vasopressors despite adequate IV fluid resuscitation. Exclusion criteria were patients who had received a previous continuous infusion of vasopressors during this hospital admission, an ongoing requirement for systemic steroid treatment (i.e., known adrenal insufficiency or regular systemic steroid therapy within the last 3 mo), end-stage renal failure, known mesenteric ischemia, Raynaud's phenomenon, systemic sclerosis or other vasospastic disease, ongoing treatment for an acute coronary syndrome, death anticipated within 24 hours or if there was a treatment limitation within place, known pregnancy, enrollment in another interventional trial that might interact with the study drugs, or hypersensitivity to any of the study drugs.
Intervention and Treatment Allocation
Enrollment, randomization, and data collection were via an online system (InForm; Oracle Corp, Redwood Shores, CA). Patients were assigned to one of two treatment groups (vasopressin and hydrocortisone or vasopressin and placebo) on a 1:1 basis with variable block randomization with two block sizes (six and four) using computer-generated random number stratified by center. The allocation sequence was prepared by an independent statistician in the Imperial Clinical Trials Unit, and it was concealed from all investigators and treating clinicians.
All patients were allocated to receive vasopressin (titrated up to 0.06 U/min) as the initial vasopressor infusion via a central venous catheter, titrated to maintain the target mean arterial pressure (MAP). A 20-unit ampoule of vasopressin (1 mL) was mixed with 49 mL 0.9% saline by the ICU bedside nurse and started at 2 mL/hr (0.013 U/min) and titrated as required to achieve the MAP target. The protocol recommended a MAP of 65–75 mm Hg, but this could be altered by the treating physician if clinically indicated. Once the maximum infusion rate of vasopressin was reached (0.06 U/min), patients received either hydrocortisone phosphate (50 mg) or placebo (0.5 mL 0.9% saline) according to treatment allocation. The hydrocortisone/placebo was administered as a 50 mg IV bolus 6 hourly for 5 days, 12 hourly for 3 days, and then once daily for 3 days, as previously reported. It could be weaned quicker than this if the shock had already resolved. If the patient was still hypotensive after the first dose of hydrocortisone/placebo, then additional open-label catecholamine vasopressors could be administered. As the patient recovered, these additional catecholamine vasopressors were weaned first and only once they were weaned off was the vasopressin weaned, that is, vasopressin was the first vasopressor infusion to start and the last to be stopped. All other treatment was at physician discretion based on the Surviving Sepsis Campaign guidelines in place at the time.
Patients could present and be recruited from any part of the hospital prior to ICU admission. Although the aim was to use vasopressin as the initial vasopressor, study drugs could not be stored in multiple locations within the hospital. Therefore, in an emergency, patients could be resuscitated using "normal" clinically prescribed vasopressors. In this situation, the patient had to be enrolled into the trial within 6 hours of commencing the vasopressor infusion. As the trial vasopressin infusion was titrated up as detailed above, the initial vasopressor infusion was weaned off as quickly as possible to maximize the vasopressin infusion rate.
Data and Sample Collection
Clinical information was recorded daily while in the ICU, and the patient was followed-up to obtain day 28 and hospital discharge status. Plasma samples for vasopressin measurement were collected once the vasopressin infusion rate was at 0.06 U/min (T0) before hydrocortisone/placebo administration if possible, 6–12 hours (T1) and 24–36 hours (T2) after the first dose of hydrocortisone/placebo was administered, and then again on day 7 if still in ICU (T3). Blood was collected into chilled EDTA tubes on ice, spun, and separated immediately and the plasma stored at –80°C until analysis. Vasopressin measurement was carried out blinded to all clinical information using a radioimmunoassay.
Statistics
The primary outcome was the difference in plasma vasopressin concentration between treatment groups. Based on data from the VASST study, to detect a 33 pmol/L difference in vasopressin levels at 6–12 hours post corticosteroid administration, assuming a SD of 45 pmol/L with a significance level of 0.05 and 80% power, 30 patients were required in each treatment group.
The difference in plasma vasopressin levels between treatment groups at T1 and T2 was analyzed on an "as treated" basis (vasopressin-corticosteroid interaction analysis). For the unadjusted analysis, the difference in mean levels at T1 is reported with a 95% CI. Adjusted analysis was carried out using regression models, with random effects incorporated to allow vasopressin levels at T1 and T2 to be jointly modeled. The secondary outcomes (clinical analyses) were analyzed on an intention-to-treat basis and include difference in vasopressin requirements between treatment groups, 28-day, ICU, and hospital mortality rates, the onset of new organ failure, and organ failure–free days in the first 28 days.
The Sequential Organ Failure Assessment (SOFA) score was used to define organ failure (score of 3 or 4 for each component) except renal failure where the Acute Kidney Injury definition of stage 3 was used. Last observation carried forward was used to impute any missing values as it is likely that the data were not collected because no change was expected by the clinician. The 3.6% of the SOFA component scores were imputed in this way, and these imputed values were sense checked against the patient's other daily scores. If normal baseline creatinine values were unknown, then these were estimated using the Modification of Diet in Renal Disease equation (for three patients).
All randomized patients who received the study drugs are included in the baseline table and safety data analysis and allocated to treatment group according to intention-to-treat. Continuous variables are summarized using the median and interquartile range, and dichotomous and categorical variables are presented in terms of frequencies and percentages. R software version 2.15.2 (http://www.r-project.org; Vienna, Austria) was used for analysis.