In this cohort study, we prospectively analyzed all patients who were consecutively admitted with a diagnosis of AHF and received treatment with NIV in an 18-bed ICU of a university hospital. The analyzed patients were those admitted between January 1997 and December 2017. The study was approved by the ethics committee of the institution.

Patients were included when the following criteria were met: a) clinical criteria: respiratory distress: presence of respiratory rate greater than 30 breaths per minute, or activity of the accessory respiratory muscles, or abdominal paradoxical breathing; physical examination: presence of bilateral crackles accompanied or not by wheezing, or auscultation of a third cardiac sound; ortopnea; severe acute or chronic on acute respiratory failure: presence of PaO2/FiO2 less than 250mmHg and/or respiratory acidosis (arterial pH less than 7.35 with PaCO2 greater than 45mmHg); b) confirmation of pulmonary congestion (at least 2 of the following): signs of pulmonary congestion on a chest imaging test; 3 or more B lines on the thoracic ultrasound in 2 thoracic areas in each hemithorax; elevation of pulmonary capillary pressure (> 18mmHg) shown by right heart catheterization; increase in total lung water by pulse contour analysis and thermodilution; echocardiographic signs of elevation of ventricular filling pressures; significant elevation of N-terminal pro-B-type natriuretic peptide above normal levels according to the patient's age. Exclusion criteria were those usually considered.19

The episode of AHF was classified according to its etiology as secondary to ACS-AMI or due to another cause. The diagnosis of ACS-AMI was established according to the current criteria at the time of admission.20–22 During the first years of the study, patients were classified as having non-Q wave myocardial infarction and Q wave myocardial infarction. These patients were subsequently reclassified as non–ST-segment elevation ACS and ST-segment elevation myocardial infarction, or indeterminate (presence of pacemaker or left branch full block). All electrocardiograms were evaluated by 2 intensive care physicians with extensive experience in coronary patients and cases of doubt being decided by the opinion of a third intensivist. Diagnosis of AMI required corroboration from the cardiologist who treated the patient in the admissions ward.

“De novo” AHF was defined as the absence of a previous history of heart failure.23

AHF patients’ treatment was carried out according to current guidelines at the time of hospital admission. The center where the study was carried out did not have a cardiac catheterization laboratory, and coronary studies were performed at the referral center 10km away, the transfer being carried out by medicalized ambulance with availability of NIV and medical and nursing personnel previously trained for this therapy. Between 1997 and 1999, in patients who needed urgent coronary angiography, that intervention was performed in the morning of the admission day (08:00-15:00) or on the next day if they were admitted after 15:00. From 2000, the availability of urgent coronary angiography was extended to 24hours a day, every day of the year. Coronary angiography studies were defined as immediate or primary when performed at the time of STEMI diagnosis, or delayed if performed after at least 24hours of diagnosis. The treatment of AHF was mainly based on the use of vasodilators and loop diuretics, both in continuous perfusion. All patients received intravenous morphine, to decrease anxiety and dyspnea, and improve adaptation to NIV, except for those with an impaired level of consciousness. The use of dobutamine or levosimendan was restricted to patients not responding to the initial treatment and showing a depressed left ventricular ejection fraction (LVEF) on cardiac ultrasound. All patients underwent bedside echocardiography in the first 24hours of stay in the ICU.

NIV was performed using specific noninvasive ventilators (BiPAP ST-D ventilator and VISION ventilator from Respironic Inc, USA; and V60 ventilator from Phillips Respironic, USA), using bilevel positive airway pressure, as previously published.19 The interface chosen in the first place was facial mask, with the nasal or the total facial mask being used when the patient showed interface-related complications. NIV was initiated with a positive inspiratory airway pressure of at least 12 cmH2O, with increments of 2 to 3 cmH2O if tolerated by the patient and according to clinical response, not exceeding 25 cmH2O. Positive expiratory pressure was initially 5 cmH2O, increasing by 1 to 2 cmH2O to improve hypoxemia or patient comfort. If the continuous positive airway pressure (CPAP) mode was used, due to intolerance to the bilevel mode, a continuous positive pressure of 7 cmH2O was used, with the possibility of rising up to 15 cmH2O. The inspired oxygen fraction (FiO2) administered was that necessary to reach an SpO2 of 92%. Since 2012, high-flow oxygen therapy through nasal cannula has been used in patients not tolerating NIV or to facilitate weaning from it, alternating the 2 respiratory therapy devices.

NIV failure was assumed when the patient showed a deterioration of gas exchange parameters, or worsening of signs and symptoms of respiratory failure, despite optimization of NIV settings, leading to intubation or death during the ICU stay or during the first 24hours in the ward after discharge from the ICU, without recovery of the respiratory process motivating the need for noninvasive ventilatory support. In case of NIV failure, the patient was intubated and connected to invasive mechanical ventilation, according to previously published criteria.19

At the beginning of NIV treatment, demographic, clinical and laboratory variables were collected. As in other studies, to quantify the patients’ comorbidity, the Charlson comorbidity index was calculated, the degree of multiorgan dysfunction was assessed through the Sequential Organ index Failure Assessment (SOFA), calculated daily, and disease severity was obtained by calculating the Simplified Acute Physiology Score II (SAPS II) index.14,19 Anemia was defined as the presence of hemoglobin less than 12g/L in women and less than 13 in men. The location of the ACS-AMI was classified as anterior, inferior or indeterminate, by means of the alterations shown in the electrocardiogram and echocardiography. ACS-AMI of lateral location were included in the inferior or anterior group, depending on the involvement of these cardiac regions. The determinations of cardiac injury markers changed over time; the level of creatine phosphokinase fraction Mb was used for the first 5 years of the study and after that troponin I. The patients were monitored up to 1 year after admission, to analyze mortality and eventual readmissions.

Qualitative variables are shown as absolute and relative frequencies, and comparisons between them were made using the Pearson chi-square test or Fisher test. Quantitative variables are expressed as mean±standard deviation or median [interquartile range], and comparisons between independent groups were made by Student t test or Mann-Whitney U test if the variable did not follow normal distribution. The Student t test was used for related data to analyze the quantitative variables measured before and after NIV initiation. The association measures analyzed were odds ratios (OR) with their 95% confidence intervals (95%CI). The changes per year was assessed through survival analysis and comparison between groups by Cox analysis. A paired propensity analysis was performed, “nearest neighbor” mode, with a 1:1 ratio, using the matching variables: age, sex, SAPS II, initial SOFA, Charlson index, and the presence of nonintubation order. All analyses were performed by bilateral contrast, and a P value less ≤ .05 was considered significant. The analysis was performed using the SPSS 25.0 program (IBM, Armonk, NY, United States) and R version 3.4.0 (Copyright 2017 The R Foundation for Statistical Computing Platform).

The main characteristics of the patients are shown in table 1. There was a slight predominance of male sex (53.7%), with a larger number of men in the ACS-AMI group. However, neither age (mean of 74 years) nor comorbidity index differed between the 2 groups. SAPS II score was slightly higher in the group without AMI. In patients with ACS-AMI, “de novo” AHF predominated. Among antecedents, hypertension and previous atrial fibrillation were more frequent in the group without infarction, while family history and previous ischemic heart disease were more common in the ACS-AMI group.

In the ACS-AMI group, 186 patients (46.2%) had STEMI, 207 (51.4%) non–ST-elevation acute myocardial infarction and 10 (2.5%) were undetermined. The most common infarction location was anterior (320 cases, 79.4%), followed by inferior (75 cases, 18.6%) and 8 cases (2%) were of undetermined location. Coronary angiography was performed in 369 patients (91.6%); in 133 (36%) angiography was immediate. Another 20 (13.1%) patients received systemic fibrinolysis. The vessel responsible for ACS-AMI was the left main artery in 43 cases (11.7%), the anterior descending artery in 214 (58%), the circumflex artery in 49 (13.3%), and the right coronary artery in 63 cases (17.1%).

Respiratory, hemodynamic and neurological parameters before and after 1 hour of NIV are presented in table 2. The initial values showed higher blood pressure and heart rate in the group without ACS-AMI with lower pH and higher PaCO2, as well as a worse neurological status. Bicarbonate and arterial lactate values also showed differences between the 2 groups, with a lower level (for bicarbonate) and a higher level (for lactate) in the ACS-AMI group. Although these differences were statistically significant, they were not clinically relevant. In both groups, all variables improved significantly at the time of ventilatory therapy (all P values <.001), except for arterial pH and serum bicarbonate in the ACS-AIM group and serum bicarbonate in the non–ACS-AMI group. The mean LVEF of the patients analyzed showed a greater compromise in the ACS-AMI group.

At the beginning of ventilatory therapy, a higher number of patients with ACS-AMI received CPAP (table 3). All other variables analyzed were similar between the 2 groups. Although NIV use was longer in patients without ACS-AMI, this difference was not statistically significant. Neither the number of patients with complications, nor most of them individually, showed differences between the 2 groups. Only 1 complication, pain of coronary etiology, was more frequent in patients with ACS-AMI during the period of NIV.

Neither the initial SOFA index score, nor subsequent scores during NIV or ICU stay differed between the groups (table 4). Although NIV failure (need for intubation or death during NIV with do not intubate order), mortality in the ICU, hospital and at 1 year, were slightly higher in the group with ACS-AMI, these differences were not statistically significant. There were 101 deaths (80.8%) in patients with NIV failure, and 56 deaths (6.3%) in patients with successful NIV (P <.001). The relationship between NIV failure and hospital mortality was observed both in patients with ACS-AMI (OR, 73.501; 95%CI, 33.364-161.918) and in non–ACS-AMI (OR, 57.474; 95%CI, 28.284-116.786).

After adjustment, the 2 groups of patients showed a more balanced distribution of the variables (table 5). Although complications, NIV failure, ICU and hospital mortality were more frequent in the group with ACS-AMI, these differences were not significant. The increased risk of NIV failure in the ACS-AMI group was 4.3% (95%CI,−2.3 to 2.4) and that of hospital death was 5% (95%CI, −2.3 to 2.5). The 1-year mortality (figure 2) showed an adjusted hazard ratio of 1.456 (95%CI, 0.986-2.152) for the ACS-AMI group (P=.059). Hospital readmission per year is shown in figure 3. The hazard ratio adjusted for readmission in the ACS-AMI group was 1.467 (95%CI, 0.977-2.201) (P=.062).

Figure 2.

One-year mortality. ACS-AMI, acute coronary syndrome-acute myocardial infarction.

(0.14MB).

Figure 3.

One-year hospital readmission. ACS-AMI, acute coronary syndrome-acute myocardial infarction.

(0.12MB).

This study has some strengths and weaknesses. This is a large prospective cohort study of patients with AHF followed up for 1 year that analyzed multiple clinical, physiological variables. It was performed in the “real world” and included all patients consecutively treated with NIV. In addition, the study included a propensity analysis to further specify the relationship between the presence of ACS-AMI and patient outcomes.

A weakness of this study is that this is an observational study and therefore subject to multiple biases when evaluating the results of the efficacy of the ventilatory technique. It was also carried out over a long period with the consequent changes in some of the definitions on the diseases analyzed; however, we attempted to overcome this weakness by homogenizing patients and analyzing the recorded data. In addition, the long study period led to variation in the sociodemographic, clinical and outcome characteristics of the patients, which may be analyzed in future works. This study was carried out in an ICU with extensive experience in NIV, both in cardiac and pulmonary diseases, with perfectly trained medical and nursing staff, in addition to having a wide range of devices, ventilators and interfaces, which may affect extrapolation of the results to other areas with less experience or resources. Finally, despite the huge number of variables collected, we may have omitted some important variables. Despite the limitations mentioned, we believe that the results are perfectly valid.