Blood Pressure and OXygenation Targets After OHCA

In comatose patients resuscitated from out of hospital cardiac arrest (OHCA), neurological injuries remain the leading cause of death. The in-hospital mortality is reported at 30-50%, and the total mortality, although improved substantially over the last decade, remain to be significant, in most countries at up to 90%. An adequate blood pressure must be maintained in the post-cardiac arrest patient i order to optimize neurological recovery and avoid further brain injury. Blood pressure targets in post-resuscitation guidelines are based on limited clinical evidence. Furthermore registry and clinical data suggest a u-shaped relationship of outcome with levels of oxygen supplementation. Blinded, randomized, clinical trials addressing specific blood pressure- or oxygenation-targets during the post-resuscitation care, have not been performed.

The current trial addresses strategies for neuroprotection using a 2-by-2 design of two different target blood pressure levels and two different oxygenation levels.

Intervention:

• 'Low-normal MAP' (appoximately 63 mmHg) vs. 'high-normal MAP' (approximately 77 mmHg) (double blind intervention) and
• Low-normal oxygenation (9-10 kPa) vs. high-normal oxygenation (13-14) kPa (open label).
• As a subordinate study, the patients will be randomized 1:1 to active fever-control with an automated feedback temperature control device for 72 hours or to 36 hours following return of spontaneous circulation.

Design: National collaborative, randomized clinical trial randomizing 800 comatose out-of-hospital cardiac arrest patients undergoing targeted temperature management (TTM) to the specified interventions.

The investigators have planned the following sub-studies:

Sub-study 1: Devopment and validation af a method for double blinded allocation to different blood pressure targets.

Hypothesis: It is possible to develop a method for double blinded allocation of patients to different blood pressure targets in clinical trials.

Sub-study 2: Assessment of different blood pressure targets and relation to renal function during TTM.

Hypothesis: Different blood presure goals will affect biomarkes of renal function after cardiac arrest.

Sub-study 3: To investigate the hemodynamic profil in relation to different blood pressure targets after cardiac arrest.

Hypothesis: Blood pressure and vassopressor-doses are related to hemodynamic parameters, such as systemic vaskular resistence index and cardiac index.

Sub-study 4: To investigate the hemodynamic profil in relation to different oxygenation targets after cardiac arrest.

Hypothesis: Lower oxygenation targets are related to higher pulmonary vascular resistance.

Sub-study 5: The prognostic value of automated videobased assessment of pupillary dilatation and reaction to light. Derivation and validation of relevant cut-off for introducing pupillomtry as part of the prognostication

INTERIM ANALYSIS There will be an independent DSMC arranging an independent statistician to conduct primarily a blinded interim analysis at time points of their choosing. The DSMC will be able to request unblinding of data coordinated by the data managing agency. An interim analysis is planned after inclusion of 200 and 400 patients.

For the BP intervention, a blinded interim analysis of vasorepressor need and recorded blood pressures is planned after 50 patients, to monitor blinding of treatment allocation and that a clinically relevant blood pressure separation between groups is achieved. Vasopressor needs in terms of vasopressor need in a variance component model is expected to differ. New sites will be monitored for these factors after inclusion of 50 patients.

EARLY STOPPING CRITERIA After an interim analysis the DSMC may suggest to the steering committee that the trial should be stopped early. No specific criteria to guide the DSMB will be put forward.

ACCOUNTABILITY PROCEDURE FOR MISSING DATA/POPULATION FOR ANALYSIS Trial sites will be asked to complete all CRFs and other forms if missing data is found in the electronic database. Missing data will be reported in the publications. More than 5% missing data will result in multiple imputation with the creation of 5-10 imputed datasets to be analysed separately and the aggregated into one estimate of intervention effect on the primary and secondary outcomes. Analyses will be performed according to the modified intention to treat principle with patients lost to follow up included in the denominator.

SUBGROUP ANALYSIS AND DESIGN VARIABLES Subgroups will be analysed according to pre-defined design variables: over or under median age, shockable rhythm, gender, the presence of shock at admission, diagnosed AMI and time from arrest to ROSC. Difference in intervention effect estimates according to subgroup will be declared exclusively based on a statistically significant test of interaction.

DIRECT ACCESS TO SOURCE DATA/DOCUMENTATION The principal investigator and the site investigators will permit monitoring, audits, review of ethical committees and regulatory authorities direct access to source data and documentation, blinded to treatment allocation.

DATA HANDLING AND RECORD KEEPING Individual patient data will be handled as ordinary chart records and will be kept according to the legislation (e.g. data protection agencies) of the countries of each health system. The study database will be stored for 15 years and anonymised if requested by the relevant authorities.

Danish legislation regarding the respect for patients physical and mental integrity and rights will be respected, Approval for storing data relevant to the trial, including potentially sensitive information has been approved by the relevant authorities.

QUALITY CONTROL AND QUALITY ASSURANCE A monitoring plan will be published before start of the trial. The monitoring will include: inclusion and absence of exclusion criteria, consent obtained in all patients.

All trial sites will be provided with sufficient information to participate in the trial. The site investigator will be responsible for that all relevant data is entered into the electronic CRFs. The CRFs will be constructed in order to assure data quality with predefined values and ranges on all data entries.

STATISTICAL METHODS The combined primary outcome will be reported as proportional hazard of experiencing one of two endpoints (death or poor neurological status at hospital discharge), differences tested with a log rank test. Other proportions are expected to be normally distributed; therefore a t-test is applied. Survival analyses are performed using proportional hazard models, survival is adjusted for site.

Furthermore pre-specified analysis of interaction for design variables: sex, age (median), time to ROSC (median), shockable rhythm, STEMI, pre-existing hypertension, pre-existing chronic obstructive pulmonary disease.

SIGNIFICANCE A two-sided significance level of 0.05 will be applied to all endpoints. No adjustment for the factorial design is made, as no interaction is expected.

SAMPLE SIZE ESTIMATION Sample size estimation is based on blinded BP target allocation and on the assumption that no interaction of the two interventions exist.

The combined primary outcome is time to death or hospital discharge in a state of CPC 3 or 4. The investigators are planning a study with 400 subjects in each group, an accrual interval of 48 months, and additional follow-up after the accrual interval of 3 months. Prior data indicate the 6 months mortality is 33% overall. Assuming a mortality of 28% in the superior group compared to 38% in the inferior groups the investigators will need to include 732 patients in total or 846 patients in total to achieve a power of 0.8 and 0.9 respectively. The Type I error probability associated with this test of the null hypothesis that the experimental and control survival curves are equal is 0.05.

Loss of final measurement is expected but from the experience from previous trial the number of missing follow-up assessments is small (<5%) and will not result in an increase of the number of patients needed.