The ELISABETH trial (NCT03683212) is a stepped wedge clinical trial in France.[27] The primary objective of this study is to assess the change in the early (one month) morbidity and mortality of AHF in elderly with the implementation of an early and comprehensive care bundle in the ED.
Experimental plan of the stepped wedge design
In this stepped wedge clinical trial, patients will be recruited in 15 EDs in France, academic and non academic, rural and urban (table 1). All clusters will begin the trial with the “control period” where included AHF patients will be routinely managed by the emergency physicians. After a first step of two weeks, every two weeks, one center will randomly be assigned to switch to the “intervention period” where the ELISABETH care bundle will be implemented. After 32 weeks, all centers will be in the “intervention period” for the four remaining weeks of the trial (Fig 1).
We decided to choose this design for the following reasons:
- As we implement a new protocol, there is a risk of contamination. An emergency physician, who would have already treated patients via the care bundle protocol, would be subsequently influenced by this trial, and could have difficulty to provide the former “standard of care”. Therefore, a randomization at the patient level or a cross-over design would induce bias through contamination.
- The present ELISABETH trial focus on a severe condition, in EDs that are often busy places, therefore the need for a randomization at the patient level could be an impediment to inclusion, and therefore limit our ability to recruit consecutive patients
- A cluster stepped-wedge design prevents contamination that could arise from a cluster cross-over design, as centres will first be allocated to standard care, before implementing the intervention. Furthermore, a stepped wedge design would also prevent a potential “period effect” that could have resulted from a simple before/after design.
Selection of participants
The ELISABETH trial focuses on elderly acute heart failure in the ED. Therefore, all patients aged 75 and over, affiliated to French social security with a diagnosis of AHF in the ED, defined by the association of:
- At least one of the following symptoms: acute or worsening dyspnea, orthopnea
- AND at least one of the following:
- Bilateral pulmonary rales or peripheral Edema
- Signs of pulmonary congestion on chest radiography or cardiac echography
- Increased natriuretic peptides (BNP or NT pro BNP)
A written informed consent signed by the patient will be thought prior to inclusion. If the patient is unable to consent, then the physician will sought consent from a trustworthy person, family member or close relative. If none are available, the physician will be able to proceed to an emergency inclusion and then the written informed consent will be signed by the patient (if need be by trustworthy person, family member or close relative) as soon as possible (article L1122-1-2 of the French Public Health Code).
Patients are excluded if they have any of the followings:
- Other obvious cause of acute illness (severe sepsis, ST elevation Myocardial infarction)
- Systolic blood pressure less than 100mmHg
- Any contra-indication to nitrates (severe mitral or aortic stenosis, or severe aortic regurgitation)
- Known chronic kidney injury on dialysis
- Time from ED entrance to inclusion > 6h,
- Patient under legal protection measure (tutorship or curatorship) and patient deprived of freedom
Trial Objectives and outcomes
The main objective of the present trial is to compare the efficacy of an early and comprehensive management strategy of AHF in elderly patients to the usual care on morbi-mortality at 30 days. Our primary endpoint is the number of days alive and out of hospital at 30 days after the index ED visit. This endpoint is considered as relevant by the group of experts of the European Society of Cardiology.[28] In their consensus paper, the experts stated that although mortality should be captured, repeated hospitalizations should also be recorded. Especially in elderly patients, where the rate of readmission to the ED and rehospitalisation is elevated: up to 40% of heart failure admissions to the hospital are actually repeated admission for recurrence of symptoms within 30 days of a previous AHF event. [3,6] We chose the timeframe of 30 days because a shorter timeframe would not catch recurrence and morbidity, and a longer timeframe would catch events that are more likely linked to chronic morbidity of the patients than to the AHF syndrome.[22,28,29]
The primary endpoint (days alive and out of hospital at day 30) will be measured at the end of the 30 days follow up period, either by hospital visit, phone interview, and medical chart review. Vital status, date of death and date of discharge will be collected.
A death during the follow up period will correspond to 0. An ED visit will correspond to “one day” at the hospital. For example, a patient not admitted (at day 0), with no return visit to the hospital, and alive at day 30 will have 30 days alive and out of hospital. A patient who is admitted (at day 0) and stays 8 days in the hospital before being discharged and have no readmission and no return visit to the ED would have a “22 days alive and out of hospital at 30 days”. A patient that is admitted and die at 13 days, either at home or in hospital will have 0. A patient that is admitted for 10 days, discharged home for 5 days then admitted at day 16 for 15 days will have 5 days alive and out of hospital (namely day 11, 12, 13, 14 and 15).
The secondary endpoints include:
- 30 day all-cause mortality
- 30 day cardio-vascular mortality
- Hospital readmission at 30 days
- Length of in hospital stay truncated at 30 days
- Changes of more than 2 fold in creatinine level from inclusion to day 30 or to discharge whichever comes first.
Creatinine will be measured at day 0 in the ED, and at discharge day or day30, whichever comes first.
Description of the intervention
This is an intervention study, where the intervention comprises the application of recommendations and guidelines for the management of AHF, ACS and infection.
Control period: Acute heart failure standard therapy:
- Treatments are given at the discretion of the treating emergency physician
The guidelines and standard of care will be recalled to the emergency physicians at the beginning of the trial in each center when the control period will start.
Intervention period: Early intensive care bundle:
The care bundle comprises a list of items to follow and tick on a handover checklist (Figure 2 & 3) within 4 hours of ED management:
- a) Treatment of the congestion: (international guidelines and recommendations [10,11])
- 40mg of Intravenous (iv) furosemide (or usual daily dose) if not already given pre-hospital.
- IV nitrates given in boluses of 3mg every 5 minutes. After one hour of bolus titration, then continuous infusion with an hourly dose of at least half of total given during the first hour of nitrate. Blood Pressure (BP) will be monitored every 5 minutes during the titration (then hourly), and nitrates will be discontinued if BP drops <100mmHg.
- b) Treatment of precipitating factors :
- Administration of antibiotic therapy (accordingly to local guidelines amoxicillin and clavulanic acid in most cases) if presence of at least two of the following: Fever > 38°C, leucocytes > 12 000 G/L, radiological signs suggestive of lower respiratory tract infection or elevated CRP or PCT,
- Administration of dual antiplatelet therapy and transfer to cardiac intensive care unit if presence of at least 2 of the followings: chest pain, ischemic signs on ECG, elevated troponin concentration or change in troponin concentration. These patients will be transferred for coronary angiography if indicated by the cardiologist, as recommended.[30]
- In case of atrial fibrillation: administration of heparin, heart rate control strategy (digoxin or amiodarone as indicated) to reduce heart rate under 100 bpm, early admission to a cardiac intensive care unit if elevated troponin associated.
- c) NIV if respiratory distress with hypercapnia and pH < 7.35 in absence of contra indication.[11]
- d) Preventive LMWH if no pre-existing anticoagulation therapy.[11]
All treatment will be initiated in the ED, and their continuation or discontinuation will be evaluated by the treating physician during the subsequent hospital stay.
Statistical Analysis
No interim analysis is planned. Analysis will be performed at the end of the study after data review and freezing of data base. Analyses will be performed using SAS® software (version 9.3 or updated version). Principal analysis will be realized according to the ITT principle.
Baseline patient’s characteristics will be considered at both with the cluster (center) and patient level. For the center level, characteristics at the beginning of the study will be described (there are no expected change between the two periods for cluster characteristics). Baseline characteristics of patients will be described globally and according to the period. Continuous variables will be summarized using descriptive statistics, i.e number of subjects, mean, standard deviation (s.d), median, inter quartile range, minimum and maximum. Qualitative variables will be summarized by frequency and percentage.
The number of days alive and out of hospital will be calculated based on date of admission, vital status, date of death and date of discharge will be collected.
This primary endpoint will be analysed using a linear regression mixed model with a random effect for each cluster, considered fixed effects will be: strategy and, for the stepped wedge design, time representing each step. In case of non-normality distribution of the interest variable, a transformation could be realized.
All causes mortality at 30 days, cardiovascular mortality at 30 days and hospital readmission at 30 days will be compared between groups by using Pearson's chi-square test or Fisher exact test.
If possible, generalized linear regression mixed model with Poisson distribution will be performed. If the number of events is sufficient, generalized linear regression mixed model using logit link will be performed. The length of stay in hospital in days will be compared between the two periods by using Student t-test or Wilcoxon rank-sum test as needed. If possible, a linear regression mixed model will be performed. A random effect for each cluster will be considered and considered fixed effects will be: strategy and, for the stepped wedge design, time representing each step. In case of non-normality distribution of the interest variable, a transformation could be realized. Percentage of patients with a change of more than 2 fold in creatinine between inclusion and 30 days will be compared between groups by using Pearson's chi-square test or Fisher exact test. If possible, generalized linear regression mixed model with Poisson distribution will be performed. If the number of events is sufficient, generalized linear regression mixed model using logit link will be performed.
Second analysis will be performed on the per protocol population.
All tests will be performed with an alpha set at 5%.
Sample size calculation
From our previous cohort, the mean number of days alive and out of hospital at 30 days was 14±9. To be clinically relevant, we estimate that the new approach should increase this endpoint of 3 days at least (a relative increase of 20%). With a power of 80% and alpha=5%, 283 patients are needed to be included. Considering the stepped-wedge design and after specification of the following elements: 15 clusters, ICC=0.0001, the design effect is estimated at 1.609, increasing to 454 subjects who needs to be included. Taking into account 10% of non-evaluable patients, it is necessary to include 500 patients - 2 per cluster for each 2 week period.