The design of the ETIFIC trial has been previously published.32 The trial was a 6-month, 2-arm, parallel, multicenter randomized controlled open-label trial carried out in 20 mostly tertiary, but with some secondary hospitals with HF units, in 10 Spanish autonomous communities.

Patients with HF, NYHA II-III and LVEF ≤ 40% were included after hospitalization in a cardiology ward. We excluded patients with elective surgery, contraindication to BB or already receiving 100% of the target dose or on the maximum tolerated dose, need for home or end-of-life care, or inability to care for themselves.

An active supervision system for the recruitment of eligible patients was established in each hospital, with centralized randomization in Galdakao Hospital, following a 1:1 intervention/control ratio, in blocks of 4 patients each and stratified by hospital, using computer generated tables. The randomization list generated by this process was concealed and safeguarded. A 4-month titration period was established, and a 6-month follow-up period after inclusion.

A safety and clinical adjudication committee, blinded to the group assignment, participant and hospital, monitored the safety of the research and evaluated all AE. In addition to the statement of the researchers from each site, AE monitoring included a review of all prespecified variables possibly associated with titration, and AE resolution at 4 months, as well as active supervision of admissions, mortality, and clinical reasons for loss of follow-up in all participants. All events were blindly evaluated by at least 4 members of this committee and, if there were discrepancies, by a fifth member. The final decision was made at least by 3 members. The study was approved by the Clinical Research Ethics Committee of the Basque Country and the Ethics Committees of each hospital (n=20). Written informed consent was requested. This study was conducted in accordance with the principles of the Declaration of Helsinki.

The protocol of the 2 randomized groups, described in the previously published design article32 is briefly summarized:

The primary objective was to compare the achieved beta-blocker mean relative dose (% relative to target dose) in the HF nurse and HF cardiologist groups after 4 months of titration. The % of target dose was defined following the target dose recommendations of the ESC HF guidelines.2

The secondary objectives were to compare the following between the 2 groups: a) the mean relative doses of ACE inhibitors, ARB and MRA, after 4 months of titration; b) the percentage of AE attributable to dose changes over the 4 months of titration; c) variables influencing target dose achievement; d) rates of mortality and readmissions at 6 months after the start of titration; and e) changes in LVEF, NYHA class, 6-minute walk distance, NT-proBNP levels and quality of life scores throughout the study according to group allocation.

A noninferiority hypothesis was set. Variables are shown in the design article.32

For the hypothesis that the relative dose of BBs would reach 52% by the end of the study and using an equivalence margin of 7%, we would need 157 patients per group for an alpha level of significance of .05 and a statistical power of 80% (beta of 0.80). Hence, estimating that as many as 20% of patients would be lost to follow-up, we needed to recruit 314 patients per group to test the main hypotheses.32

The analysis was performed on an intention-to-treat basis. Both the Student t test (or the nonparametric Wilcoxon test if continuous data were not normally distributed) and the chi-square test (or Fisher exact test) were used to compare the baseline sociodemographic and clinical characteristics of the patients in the 2 groups. The effect attributable to the intervention was estimated by comparing the differences in the relative dose of BB (primary endpoint of the study), ACE inhibitors, ARB and MRA reached between the groups, assessed at 4 months after the start of titration, and the 95% confidence interval (95%CI) was calculated. We performed multivariate analysis for the primary and secondary endpoints as predefined in the original study design.32 The model was adjusted by baseline dose, age, educational level, baseline heart rate, amiodarone and the number of visits of the professional who titrated the target dose/additional visit; these were established as relevant related factors with a possible effect on dosing, based on a review of the literature. All variables with P <.20 were included as explanatory variables in a multivariate model, with relative dose as the response variable. The effect of time was estimated in 2 repeated measurements for each participant, using mixed linear regression models with fixed effects (time, intervention, interaction between time and intervention) and random effects (specific effect of each participant and center at the reference level and the effect of time). These models took into account the longitudinal structure of the 2 repeated measurements, as well as the hierarchical and multicenter structure of the data. All statistical analyses were performed, using SAS System v 9.4, with statistical significance set at P <.05.

Survival analysis was estimated using Kaplan-Meier tables and the survival rate of each group was compared using the log-rank test.

See the CONSORT 2010 reporting checklist of the supplementary data.

A total of 824 patients with de novo HF were evaluated in 20 hospitals (2015- 2018). We excluded 504 patients, the main causes being elective surgery and already taking the maximal BB dose (figure 1). We included 320 patients, of whom 164 were randomized to the HF nurse group and 156 to the HF cardiologist group. Finally, 289 patients were analyzed at 4 months, 144 vs 145, and 274 at 6 months, 136 vs 138, respectively, due to losses to follow-up.

Figure 1.

Patient flowchart. HF, heart failure; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.

(0.93MB).

Patient characteristics were generally well-balanced between the 2 groups, although the HF nurse group had lower estimated glomerular filtration rate (eGFR) and a higher score (worse) in the Minnesota living with HF Questionnaire (table 1 and table 1 of the supplementary data shows loss to follow-up causes and table 2 of the supplementary data shows other baseline characteristics).

The HF nurse group achieved a BB relative dose and target dose that were not inferior to the HF cardiologist group, with statistically significant differences favouring HF nurses. The number of visits, which was higher in the HF nurse group than the HF cardiologist group, was associated with a higher dosage. However, after adjusting the multivariate analysis by this variable, among others, we found that HF nurses still achieved better results (figure 5).

Figure 5.

Take home figure. ETIFIC is the first multicenter randomized trial to demonstrate the noninferiority of drug titration by HF nurses vs drug titration by HF cardiologists. ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blocker; BB, beta-blockers; CV, cardiovascular; HF, heart failure; HFC, heart failure cardiologist; HFN, heart failure nurse; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor blocker.

(0.37MB).

The relative doses of both ETIFIC groups were lower than the weighted average achieved in RCT of BB (HF nurse −6%, HF cardiologist −21%), as well as in those measuring the target dose (HF nurse −10%; HF cardiologist −33%). These differences could be explained by the characteristics of the ETIFIC patients, all recruited after HF admission, the lower mean systolic blood pressure and heart rate, and the greater presence of diabetes, respiratory disease and smoking compared with patients in RCT.

The target doses achieved in 2 RCTs of HF nurse titration programs were close to the ETIFIC HF nurse group and higher than those in the HF cardiologist group, although these trials had small sample sizes.22,29 A RCT with joint cardiologist-nurse titration achieved lower relative doses than both ETIFIC groups. An exercise program achieved a relative dose close to the HF nurse group (see references of the supplementary data[BB RCT, 1-22]).

Previous observational studies showed wide heterogeneity in their results (relative dose 33%-63% and target dose 12%-58%).6–9,11–13,18 Those who reported having a HF nurse showed better results.

The relative ETIFIC HF nurse dose was 8% higher than an observational study dose with HF-clinic/HF-nurse18 and 23% higher than those who did not report having a HF nurse.7,9 The target dose was 15% higher than the average weighted observational studies with a HF nurse11,18 and 30% higher than the average of those who did not have a HF nurse.6,7,12,13

The HF clinics and HF nurses, as in ETIFIC, guarantee access to specialized health professionals and a sufficient number of consultations for titration, a key aspect of our study.

In contrast, the relative ETIFIC HF cardiologist dose was 7% lower than that in an observational study with HF nurses18 and 8% higher than that in those without HF nurses.7,9 The target dose was also 8% lower than that in observational studies with HF nurses8,11,18 and was 7% higher than those without HF nurses. 6,7,12,13

In addition, the percentage of patients within the target dose of the HF cardiologist group was almost double than the observed in the Spanish long-term HF Registry6 cohort (same country, similar health professionals). This suggests that a controlled environment such as a randomized trial, with a specific titration project, a protocol and assignment of a responsible professional, can improve results, as previously stated.8,11,13,14

The ETIFIC patient comorbidity profile was close to those in previous observational studies, reinforcing its applicability, although the mean age was 8 years lower because patient inclusion was strictly limited to those with de novo HF.6–9,11–13,18

However, the prescription of the 3 drug groups (BB, ACE inhibitors/ARB/ARB-neprilysin inhibitor, MRA) was much higher in ETIFIC (81%/79%) than in RCT and observational studies,6–9,11–13,18 mainly due to the low prescription of MRA in previous studies (RCTs: 11%-41%; observational studies: 15%-74%). This further enhances the feasibility of dosage improvement in a context of optimal prescription, as shown in ETIFIC (see references of the supplementary data [BB observational studies 23-34]).

Noninferiority was also confirmed in the ACE inhibitor dose achieved by HF nurses vs HF cardiologists with statistically significant differences favoring HF nurses.

As with BB, the 2 groups achieved doses lower than RCTs of ACE inhibitors and ARB. The difference could be explained by the lower baseline systolic blood pressure, the low prescription of MRA and BB (these studies were performed many years ago) and withdrawals in RCTs (up to 32%).

The ACE inhibitor/ARB doses in previous observational studies also showed greater heterogeneity.6–8,11–13 Systolic blood pressure was higher than in ETIFIC. As in ETIFIC, studies that reported having HF nurses8,11 achieved target doses twice as high as those without HF nurses,12,13 except one.

Regarding MRA, the 2 ETIFIC groups achieved doses 13% lower than those in the RCTs. It is difficult to compare observational studies, given the low prescription of MRA compared with ETIFIC (see references of the supplementary data [RCT: ACE inhibitors 14,15,35,36 and ARB 37-39, MRA 40,42; observational studies 23, 26-31]).

The number of HF nurse visits in ETIFIC was not higher than in other HF-clinic/HF nurse studies, and therefore could be considered adequate to ensure a careful titration process.

This study also demonstrated noninferiority in terms of mortality and hospital admissions. In addition, there were fewer total and HF-related hospitalizations in the HF nurse group, probably related to a higher BB and ACE inhibitor dosage, number of visits, and application of the flexible diuretic regimen.

Six-month HF hospitalization and mortality in both ETIFIC groups were lower than those reported in the literature, probably due to therapeutic optimization and inclusion of de novo HF patients only.

Noninferiority was also confirmed in the 6-month clinical outcomes, demonstrating major improvement in LVEF, NT-proBNP, NYHA, 6-minute walk, test and quality of life without significant differences between both groups.

The improvement in mean LVEF after optimization in both ETIFIC groups (15%) exceeded that achieved in the RCT of BB (6%), ACE inhibitors (2.1%), ARB (4.3%) and MRA (3.4%), considered individually. However, it was closer to observational studies in which the optimization of drugs was evaluated grouped as follows: 22% to 70% of patients showed an improvement in LVEF ≥ 10% and 51% of patients exceeded LVEF 35%. This reinforces the previously described practices2 of reevaluating LVEF after optimization, prior to contemplating therapies such as cardiac defibrillators or cardiac resynchronization.

The increase in meters covered in the 6-minute walk test in both ETIFIC groups,> 50 meters after the titration process, could be associated, according to the BIOSTAT study, with a significant reduction in HF hospitalization and mortality.

The significant deterioration in baseline quality of life, similar to that seen in previous studies after hospital admission, and marked improvement at 6 months, similar to that in studies after optimization in HF clinics, showed the benefit of this process.

The specific study of dosage based on age or sex and the in-depth study of factors associated with improvement of LVEF, quality of life, and the 6-minute walk test go beyond the scope of this study and could be the subject of future research (see references of the supplementary data [visits 43; serious AE 43-46; LVEF 1,2,4-6,8-11,16,18, 36,37,47-84; the 6-minute test 85; quality of life 26,86]).

The external validity was represented by 20 hospitals and 10 different health systems. The participants were closer to the real-world population than the highly selected participants in randomized drug trials in terms of their comorbidities. However, the external validity of the study is limited by the implementation of ETIFIC in a single country. Other limitations of the ETIFIC trial are its recruitment characteristics, inclusion of de novo HF patients exclusively, relatively young patients and the low percentage of women and patients with ischemic heart disease. Substudies of older patients in the ETIFIC trial and sex differences remain the focus for further research. Moreover, further similar studies on the implementation of this protocol for home care patients, such as the intervention of HF nurses at home with the support of cardiologists, or nonface-to-face interventions, or for patients admitted to other hospital areas, or ambulatory patients, could be designed to assess this vulnerable population.

Organizing care with a HF nurse responsible for titration can help to improve guideline implementation. ETIFIC requirements for HF nurses, protocol and measurements could also help to enhance quality indicators of the titration process.