Abstract
Background: This study investigated whether the chronic use of adaptive servo-ventilation (ASV) reduces all-cause mortality and the rate of urgent rehospitalization in patients with heart failure (HF).
Methods and Results: This multicenter prospective observational study enrolled patients hospitalized for HF in Japan between 2019 and 2020 who were treated either with or without ASV therapy. Of 845 patients, 110 (13%) received chronic ASV at hospital discharge. The primary outcome was a composite of all-cause death and urgent rehospitalization for HF, and was observed in 272 patients over a 1-year follow-up. Following 1:3 sequential propensity score matching, 384 patients were included in the subsequent analysis. The median time to the primary outcome was significantly shorter in the ASV than in non-ASV group (19.7 vs. 34.4 weeks; P=0.013). In contrast, there was no significant difference in the all-cause mortality event-free rate between the 2 groups.
Conclusions: Chronic use of ASV did not impact all-cause mortality in patients experiencing recurrent admissions for HF.
Over the past few decades, significant advances have been made in the treatment of chronic heart failure (HF).1–3 Nevertheless, the number of patients with HF has increased markedly with population aging, particularly in Western countries.4 Therefore, alternative treatment strategies aimed at enhancing prognosis and quality of life (QOL) are needed for these patients.5 Given that impaired breathing, including at rest and exercise-related shortness of breath, plays a pivotal role in HF pathophysiology, non-pharmacological ventilatory support could be beneficial in these patients.
Adaptive servo-ventilation (ASV) is a recently developed non-invasive positive-pressure ventilation (NPPV) technology that effectively mitigates airway congestion, reduces cardiac workload, and alleviates sleep-disordered breathing (SDB) by delivering servo-controlled inspiratory pressure support in conjunction with expiratory positive airway pressure.6 The clinical use of NPPV in acute congestive states is well established, and this treatment is used routinely in emergency rooms and intensive care units.7–10 Furthermore, the portability, automatic synchronization capabilities, and non-invasive nature of ASV make it a potential long-term option for patients with severe HF experiencing recurrent respiratory distress.11
Several studies have reported positive outcomes of ASV therapy in patients with HF, including improved cardiac pump function, reduced cardiac load, and a better prognosis.12–19 Notably, ASV has been linked with correction of heightened sympathetic nerve activity in these patients.19–23 However, much of this evidence has been derived from single-institution or retrospective investigations. Moreover, the results of SERVE-HF, a large-scale randomized controlled study of the effects of ASV on SDB in patients with chronic HF, suggested an increase in cardiovascular mortality.24 The findings of SERVE-HF diverged from those of previous Japanese studies, prompting speculation that the use of an unusually high positive end-expiratory pressure to mitigate SDB and at a level above that typically used to manage pulmonary congestion in Japan may have inadvertently diminished cardiac pump function and augmented sympathetic nerve activity in that trial.25
The aim of the present study was to determine whether chronic use of ASV can reduce all-cause mortality and the urgent rehospitalization rate in patients experiencing recurrent admissions for HF.
Methods
Study Design
SAVIOR-L (Prospective Cohort Study of Adaptive Servo-Ventilation Therapy on Prognosis in Repeatedly Hospitalized Patients With Chronic Heart Failure: Longitudinal Observational Study of Effects on Re-Admission and Mortality; University Hospital Medical Information Network (UMIN) Clinical Trials Registry ID UMIN000034295) is a multicenter prospective observational study of patients who were admitted for exacerbation of HF in Japan between January 2019 and December 2020 and had a documented history of HF-related hospitalizations within the previous year.
The study was approved by the Institutional Review Board of the Public Health Research Foundation (Reference no. 18G0001) and the ethics committees at all participating sites and was conducted in accordance with the ethical principles stated in the Declaration of Helsinki. All patients provided written informed consent.
Study Population and Patient Selection
Patients aged ≥20 years with HF who met the inclusion criteria were enrolled in the study. The inclusion criteria were: (1) ≥1 hospitalization for exacerbation of chronic HF (according to the 2017 Japanese Circulation Society/Japanese Heart Failure Society guideline26) within the year preceding the date of the current hospital admission; (2) able to attend outpatient visits after hospital discharge, with the flexibility of choosing a hospital other than the initial admitting facility; (3) willingness to undergo assessments at the designated medical institution for 1 year following discharge; and (4) able to provide written informed consent. The exclusion criteria were a diagnosis of dementia, participation in another clinical trial or blinded intervention study, and unsuitable for participation in the opinion of the attending physician.
After obtaining written informed consent for participation, the attending clinicians judged whether daily ASV was necessary as a home-based medical therapy for relief of congestion symptoms after hospital discharge on a patient-by-patient basis. Patients for whom ASV was recommended made the final decision whether to start this treatment.
Patients who elected to start ASV therapy at the time of discharge were classified as the ASV group, and those who declined or were deemed not to require it were classified as the non-ASV group. All patients were followed up until 1 year after the date of registration of the final case (i.e., for up to approximately 3 years). During the study period, patients in the ASV group could withdraw from ASV therapy and those in the non-ASV group could start ASV therapy as deemed necessary. The registration period was from January 2019 to December 2020. Follow-up was continued until the end of the research period (December 2021).
Data Collection
The MARVIN Electronic Data Capture system (version 2.6.5; EvidentIQ, Munich, Germany) was used to collect data from the medical records, including baseline demographics, the dates on which each patient was hospitalized, age, sex, height, body weight, symptoms, medications for HF, systolic blood pressure (SBP), heart rate, comorbidities (ischemic or non-ischemic heart disease, hypertension, diabetes, chronic kidney disease, atrial fibrillation, and stroke), echocardiography findings (left ventricular ejection fraction [LVEF]) and chest roentgenography, QOL (assessed using the EuroQol 5-dimensions 5-levels [EQ-5D-5L] questionnaire), and results of laboratory investigations. All these items were evaluated after stabilization of HF and registered before hospital discharge. B-Type natriuretic peptide (BNP) and N-terminal pro B-type natriuretic peptide (NT-proBNP) concentrations were assessed using the following conversion formula:27
Log[NT-proBNP] = 1.21 + 1.03 × log[BNP] − 0.009 × BMI − 0.07 × eGFR
where BMI is the body mass index and eGFR is the estimated glomerular filtration rate.
Endpoints
The primary endpoint was a composite of all-cause death and urgent rehospitalization. Secondary endpoints were cardiovascular events (HF-related death, sudden cardiac death caused by arrhythmia, urgent rehospitalization for worsening of HF, worsening of HF during hospitalization), an improvement in QOL, a reduction in the NT-proBNP concentration after 1 year of observation, and the clinical composite response (cardiovascular events, New York Heart Association [NYHA] functional class, and QOL).1
Study Procedures
The executive committee comprised members of the academic study (see Supplementary Appendix). The coinvestigators were responsible for the independent drafting and editing of all versions of the manuscript and study analyses. The executive committee selected an independent data and safety monitoring board, comprising qualified clinical scientists who were not study investigators. The independent data and safety monitoring board periodically evaluated the safety and efficacy data and made recommendations regarding continuation or modification of the study or study procedures.
An endpoint committee reviewed all potential primary and secondary endpoint events to adjudicate the endpoint designation. The committee members who adjudicated on events were independent from the study investigators.
Statistical Analysis
Continuous data are presented as the mean±SD and categorical data are presented as a frequency (percentage). Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). All statistical analyses were performed using SAS Release 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 4.1.0 or later (R Foundation for Statistical Computing, Vienna, Austria). P<0.05 was considered statistically significant.
In SAVIOR-C, the incidence of all-cause hospitalization or death was 13.7% (14/102) in the ASV group and 22.3% (23/103) in the non-ASV group.11 Based on these values, the estimated respective hazard values for the ASV and non-ASV groups after an observation period of 8 months were 0.221 and 0.379. The main analysis in the present study was based on the log-rank test for the population after propensity score (PS) matching. Assuming that the ratio of the ASV to non-ASV group in the study population was 1 : 3 after matching, with a 2-sided significance level of 5%, a power of 80%, an enrollment period of 2 years, and an observation period of 1 year, the required sample size was calculated to be 360 (ASV group, n=90; non-ASV group, n=270). Assuming that PS matching removed 70% of the non-ASV group from the main analyses, the required sample size was 990 (ASV group, n=90; non-ASV group, n=900). Assuming that some of the data would not be analyzable, the target number of patients was set at 1,100 (ASV group, n=100; non-ASV group, n=1,000).
The full analysis set comprised patients for whom any efficacy observations were made. We then selected the matched population in the full analysis set by PS matching using the following factors: age (<75 vs. ≥75 years), sex (male vs. female), underlying heart disease (ischemic vs. non-ischemic), LVEF (<40% vs. ≥40%), congestion (yes vs. no), SBP (<100 vs. ≥100 mmHg), NYHA functional class (≤II vs. ≥III), atrial fibrillation (yes vs. no), and kidney function (eGFR <45 vs. ≥45 mL/min/1.73 m2). We created the matched population using an absolute value of the standardized difference of <0.25 for each factor. We then analyzed the efficacy of ASV in this matched population. Complete case analysis was performed without compensation when values were missing.
“Time to hospitalization or death from any cause” was compared between the 2 groups in the matched population using survival curves, median survival, and annual survival rates for each group, which were estimated using the Kaplan-Meier method. CIs for median survival were calculated using the Brookmeyer and Crowley method, and those for survival rates were calculated using Greenwood’s formula.
Secondary endpoints were compared between groups using t-tests or the Wilcoxon test for continuous and ordinal categorical variables, the Chi-squared test or Fisher’s exact test for nominal categorical variables, and the log-rank test for time-to-event data.
Post Hoc Analysis According to NT-proBNP Concentration
After data analysis using PS matching, we performed a post hoc examination to adjust for the NT-proBNP concentration in the Kaplan-Meier curves for the primary endpoint, including all-cause death, dividing the data by a cut-off of 900 pg/mL (indicative of overt HF) or the median value.1 Kaplan-Meier analysis was performed for all-cause death and the primary endpoint in both the full analysis set and the PS-matched groups.
Results
Full Analysis Set
In total, 864 patients were enrolled. After 19 exclusions, the final study cohort comprised 845 patients (Table 1). There were 110 patients (83 men, 27 women) in the ASV group and 735 (457 men, 278 women) in the non-ASV group. The ASV group primarily included patients with more severe HF, as evidenced by a higher prevalence of NYHA Class III or IV and more patients with an LVEF <40%, lower mean SBP, higher mean pulse rate, more previous hospitalizations, elevated mean NT-proBNP concentration, lower mean eGFR, and increased use of tolvaptan in comparison with the non-ASV group. Furthermore, the age profile was younger and there were more men in the ASV than non-ASV group. In terms of previous use of ASV and continuous positive airway pressure, 50.9% and 7.3% of patients in the ASV group had a history of these therapies, respectively, compared with 1.2% and 5.9% in the non-ASV group (Table 1).
Table 1.
Baseline Characteristics in the Full Analysis Set
| |
Total cohort |
ASV group |
Non-ASV group |
P value |
| No. patients |
845 |
110 |
735 |
|
| Age (years) |
76.0±11.9 |
72.1±12.7 |
76.6±11.7 |
0.001 |
| Male sex |
540 (63.9) |
83 (75.5) |
457 (62.2) |
0.008 |
| Female sex |
305 (36.1) |
27 (24.5) |
278 (37.8) |
| BMI (kg/m2) |
21.8±4.0 (n=806)* |
22.6±4.6 (n=106)* |
21.7±3.9 (n=700)* |
0.061 |
| NYHA functional class |
(n=835)* |
(n=108)* |
(n=727)* |
|
| I |
85 (10.2) |
7 (6.5) |
78 (10.7) |
0.011 |
| II |
457 (54.7) |
48 (44.4) |
409 (56.3) |
| III |
228 (27.3) |
40 (37.0) |
188 (25.9) |
| IV |
65 (7.8) |
13 (12.0) |
52 (7.2) |
| LVEF (%) |
41.5±17.2 (n=829)* |
33.9±16.7 (n=107)* |
42.6±17.0 (n=722)* |
<0.001 |
| <40% |
418 (49.6) |
76 (69.1) |
342 (46.7) |
<0.001 |
| ≥40% |
424 (50.4) |
34 (30.9) |
390 (53.3) |
| SBP (mmHg) |
110.8±20.1 |
107.9±18.8 |
111.2±20.3 |
0.093 |
| Pulse rate (beats/min) |
72.1±13.7 (n=843)* |
73.6±13.5 |
71.9±13.7 (n=733)* |
0.225 |
| Etiology |
| Ischemic |
294 (34.8) |
40 (36.4) |
254 (34.6) |
0.748 |
| Non-ischemic |
551 (65.2) |
70 (63.6) |
481 (65.4) |
No. HF hospitalizations within
1 year |
1.4±0.8 |
1.7±1.0 |
1.4±0.8 |
0.002 |
| NT-proBNP (pg/mL) |
3,109.5 |
4,773.6 |
2,874.4 |
0.241 |
| Mean (range) |
(2.7–153,580.0) (n=800)* |
(2.7–153,580.0) (n=99)* |
(9.7–83,897.0) (n=701)* |
|
| eGFR (mL/min/1.73 m2) |
37.4±19.5 |
35.4±26.3 |
37.7±18.2 |
0.375 |
| Hemoglobin |
11.4±2.0 (n=843)* |
11.9±2.1 |
11.3±2.0 (n=733)* |
0.008 |
| Comorbidities |
| Diabetes |
353 (41.8) |
44 (40.0) |
309 (42.0) |
0.756 |
| Hypertension |
542 (64.1) |
64 (58.2) |
478 (65.0) |
0.167 |
| Previous stroke |
126 (14.9) |
15 (13.6) |
111 (15.1) |
0.775 |
| AF |
518 (61.3) |
69 (62.7) |
449 (61.1) |
0.834 |
| HF therapy |
| LVEF <40% (n) |
418 |
76 |
342 |
|
| Medication |
| ACEi/ARB |
276 (66.0) |
45 (59.2) |
231 (67.5) |
0.181 |
| β-blocker |
364 (87.1) |
66 (86.8) |
298 (87.1) |
1.000 |
| Thiazide |
47 (11.2) |
8 (10.5) |
39 (11.4) |
0.808 |
| MRA |
212 (50.7) |
38 (50.0) |
174 (50.9) |
0.835 |
| Loop diuretic |
380 (90.9) |
64 (84.2) |
316 (92.4) |
0.044 |
| Tolvaptan |
268 (64.1) |
58 (76.3) |
210 (61.4) |
0.017 |
| NPPV |
| Prior ASV |
45 (10.8) |
41 (53.9) |
4 (1.2) |
<0.001 |
| Prior CPAP |
24 (5.7) |
5 (6.6) |
19 (5.6) |
0.785 |
| LVEF ≥40% (n) |
424 |
34 |
390 |
|
| Medication |
| ACEi/ARB |
233 (55.0) |
18 (52.9) |
215 (55.1) |
0.858 |
| β-blocker |
296 (69.8) |
27 (79.4) |
269 (69.0) |
0.245 |
| Thiazide |
49 (11.6) |
4 (11.8) |
45 (11.5) |
0.849 |
| MRA |
164 (38.7) |
11 (32.4) |
153 (39.2) |
0.625 |
| Loop diuretic |
381 (89.9) |
31 (91.2) |
350 (89.7) |
1.000 |
| Tolvaptan |
250 (59.0) |
26 (76.5) |
224 (57.4) |
0.044 |
| NPPV |
| Prior ASV |
20 (4.7) |
15 (44.1) |
5 (1.3) |
<0.001 |
| Prior CPAP |
27 (6.4) |
3 (8.8) |
24 (6.2) |
0.467 |
Unless indicated otherwise, data are presented as the mean±SD or n (%). *For variables with missing values, the number of patients evaluated is shown. ACEi, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; ASV, adaptive servo-ventilation therapy; BMI, body mass index; CPAP, continuous positive airway pressure; eGFR, estimated glomerular filtration rate; HF, heart failure; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonist; NPPV, non-invasive positive-pressure ventilation; NT-proBNP, N-terminal pro B-type natriuretic peptide (after logarithmic transformation, reverted); NYHA, New York Heart Association; SBP, systolic blood pressure.
Patients received ASV with the settings described in Table 2. The primary reason for initiating chronic use of ASV was to mitigate pulmonary congestion at night. The expiratory positive airway pressure and pressure support settings were similar to the default settings. During the study period, 33 of the 110 patients in the ASV group discontinued ASV.
Table 2.
ASV Therapy Settings (n=110 Patients)
| Purpose of ASV therapy |
| Pulmonary congestion |
80 (72.7) |
| Sleep-disordered breathing |
20 (18.2) |
| Respiratory failure |
10 (9.1) |
| ASV therapy schedule |
| Daytime |
4 (3.6) |
| Night-time |
85 (77.3) |
| Both |
21 (19.1) |
| EPAP (cmH2O) |
| Maximum |
6.0±3.2 |
| Minimum |
3.9±1.5 |
| Pressure support (cmH2O) |
| Maximum |
8.6±2.6 |
| Minimum |
3.3±1.9 |
Data are presented as mean±standard deviation or n (%). ASV, adaptive servo-ventilation; EPAP, expiratory positive airway pressure.
In the full analysis set, event-free survival was significantly lower among patients who received ASV throughout the study (P=0.009, log-rank test; Figure 1A). Kaplan-Meier analysis revealed similar all-cause mortality curves for both groups (P=0.491, log-rank test; Figure 1B). Several factors were found to be significantly associated with the primary endpoint in the Cox proportional hazards regression analysis, including the presence of congestion, reduced renal function, and the prior hospitalization rate (Table 3). The NT-proBNP concentration also tended to be associated with the primary endpoint. To explore this association further, we stratified patients based on their NT-proBNP concentration, dividing them using a cut-off of 900 pg/mL (indicative of overt HF) or the median NT-proBNP level (1,341 pg/mL; Supplementary Table 1). The event-free survival in patients with higher NT-proBNP was similar between the ASV and non-ASV groups (Figure 1C,E). However, patients with a lower NT-proBNP concentration in the non-ASV group had better outcomes (Figure 1C,E). The event-free survival showed consistent trends in each group regardless of the NT-proBNP concentration throughout the study period (Figure 1D,F).
Table 3.
Multivariate Cox Proportional Hazards Model for the Primary Endpoint in the Full Analysis Set
| |
HR |
95% CI |
P value |
| Age (vs. <75 years) |
1.106 |
0.909–1.344 |
0.315 |
| Sex (vs. male) |
1.045 |
0.864–1.263 |
0.653 |
| Ischemic (vs. non-ischemic) |
1.086 |
0.899–1.313 |
0.392 |
| LVEF (vs. ≥40%) |
1.076 |
0.877–1.320 |
0.481 |
| Congestion (vs. no congestion) |
1.401 |
1.172–1.676 |
<0.001 |
| SBP (vs. ≥100 mmHg) |
1.151 |
0.947–1.399 |
0.158 |
| NYHA functional class (vs. ≥Class III) |
0.975 |
0.808–1.176 |
0.790 |
| AF (vs. no AF) |
0.914 |
0.761–1.097 |
0.335 |
| eGFR (vs. ≥45 mL/min/1.73 m2) |
1.341 |
1.090–1.649 |
0.006 |
| Diabetes (vs. no diabetes) |
0.947 |
0.790–1.136 |
0.559 |
| HF hospitalization within 1 year |
1.105 |
1.002–1.220 |
0.047 |
| NT-proBNP |
1.076 |
0.999–1.159 |
0.054 |
CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.
PS-Matched Population
PS matching between the ASV and non-ASV groups resulted in a cohort of 384 patients (ASV, n=96; non-ASV, n=288). In accordance with the study protocol, all covariates were adjusted to ensure a standardized mean difference of <0.25, indicating an acceptable balance (Table 4). In this PS-matched population, 23 of 96 patients in the ASV group discontinued ASV during the study period.
Table 4.
Differences in Nine Predefined Covariates After Propensity Score Matching
| Matching factor |
Matching criteria |
SMD |
| Age (years) |
<75 vs. ≥75 |
0.085 |
| Sex |
Male vs. female |
0.039 |
| Underlying cardiac disease |
Ischemia vs. non-ischemia |
0.043 |
| LVEF (%) |
<40 vs. ≥40 |
0.059 |
| Congestive findings |
Present vs. absent |
0.104 |
| SBP (mmHg) |
<100 vs. ≥100 |
0.066 |
| NYHA functional class |
≤Class II vs. ≥Class III |
0.056 |
| AF |
Present vs. absent |
0.015 |
| eGFR (mL/min/1.73 m2) |
<45 vs. ≥45 |
0.056 |
SMD, standardized mean difference. Other abbreviations as in Table 1.
In the PS-matched cohort, the median duration of follow-up was 1.6 years (interquartile range 1.0–2.0 years) and the mean duration of follow-up was 1.5±0.7 years. During follow-up, there were 272 primary outcomes, comprising 108 (28%) all-cause deaths and 201 (52%) hospitalizations for HF. The median time to the primary outcome was 19.7 weeks (95% CI 14.7–30.0) in the ASV group and 34.4 weeks (95% CI 26.1–42.1) in the non-ASV group (P=0.016). Throughout the study, the event-free survival remained consistently and significantly lower in the ASV group (P=0.013, log-rank test; Figure 2A). However, the survival curves for all-cause mortality in both the ASV and non-ASV groups showed an almost identical pattern (P=0.969, log-rank test; Figure 2B). Therefore, the disparity in time to the primary outcome likely stems from differences in hospitalization events.
We then conducted post hoc analyses to investigate potential explanations for the significant differences in time to primary outcomes between the PS-matched ASV and non-ASV groups. We categorized patients based on the NT-proBNP concentration, selecting NT-proBNP owing to its lack of normalization and the significant difference in its level between the PS-matched groups (Table 5). The ASV group continued to show higher NT-proBNP concentrations even after PS matching (Table 6). When we divided the study groups based on an NT-proBNP concentration of 900 pg/mL or the median NT-proBNP value (623 pg/mL; Supplementary Table 1), there was no significant between-group difference in the composite outcome in patients with a higher NT-proBNP concentration; however, patients in the non-ASV group with a lower NT-proBNP had better outcomes (Figure 2C,E). In contrast, event-free survival remained consistent across groups when stratified by NT-proBNP concentration (Figure 2D,F).
Table 5.
Comparison of Background Factors in Propensity Score-Matched Groups
| Background factor |
ASV group
(n=96) |
Non-ASV group
(n=288) |
P value |
| Age (years) |
73.4±12.4 |
75.6±12.3 |
0.126 |
| Male sex |
70 (72.9) |
215 (74.7) |
0.788 |
| Ischemic etiology |
36 (37.5) |
102 (35.4) |
0.714 |
| LVEF (%) |
35.0±17.4 |
36.6±15.9 |
0.454 |
| Positive for congestion |
53 (55.2) |
144 (50.0) |
0.410 |
| SBP (mmHg) |
108.3±18.8 |
108.9±19.0 |
0.774 |
| NYHA functional class ≥III |
43 (44.8) |
123 (42.7) |
0.719 |
| Positive for AF |
62 (64.6) |
188 (65.3) |
0.902 |
| eGFR (mL/min/1.73 m2) |
36.3±27.4 |
36.2±16.6 |
0.963 |
| HF hospitalization (times/year) |
1.7±1.0 |
1.4±0.8 |
0.011 |
| NT-proBNP (pg/mL) |
4,911.8±17,053 (n=85)* |
3,475.2±7,433 (n=272)* |
0.452 |
| Prior ASV |
48 (50.0) |
5 (1.7) |
<0.001 |
Unless indicated otherwise, data are given as the mean±SD or n (%). *The number of patients for NT-proBNP. Abbreviations as in Table 1.
Table 6.
Background Characteristics After Propensity Score Matching
| |
Total cohort |
ASV group |
Non-ASV group |
P value |
| No. patients |
384 |
96 |
288 |
|
| Age (years) |
75.1±12.4 |
73.4±12.4 |
75.6±12.3 |
0.126 |
| Male sex |
285 (74.2) |
70 (72.9) |
215 (74.7) |
0.788 |
| Female sex |
99 (25.8) |
26 (27.1) |
73 (25.3) |
| BMI (kg/m2) |
21.8±3.9 (n=366)* |
22.2±4.4 (n=92)* |
21.6±3.7 (n=274)* |
0.219 |
| NYHA functional class |
(n=379)* |
(n=94)* |
(n=285)* |
|
| I |
36 (9.5) |
7 (7.4) |
29 (10.2) |
0.835 |
| II |
177 (46.7) |
44 (46.8) |
133 (46.7) |
| III |
128 (33.8) |
32 (34.0) |
96 (33.7) |
| IV |
38 (10.0) |
11 (11.7) |
27 (9.5) |
| LVEF (%) |
36.2±16.2 (n=377)* |
35.0±17.4 (n=93)* |
36.6±15.9 (n=284)* |
0.454 |
| <40% |
260 (67.7) |
63 (65.6) |
197 (68.4) |
0.616 |
| ≥40% |
124 (32.3) |
33 (34.4) |
91 (31.6) |
| SBP (mmHg) |
108.7±19.0 |
108.3±18.8 |
108.9±19.0 |
0.774 |
| Pulse rate (beats/min) |
72.1±12.8 (n=383)* |
73.0±13.3 |
71.8±12.7 (n=287)* |
0.448 |
| Etiology |
| Ischemic |
138 (35.9) |
36 (37.5) |
102 (35.4) |
0.714 |
| Non-ischemic |
246 (64.1) |
60 (62.5) |
186 (64.6) |
| HF hospitalizations within 1 year (n) |
1.5±0.8 |
1.7±1.0 |
1.4±0.8* |
0.011 |
| NT-proBNP (pg/mL) |
3,817.2 |
4,911.8 |
3,475.2 |
0.452 |
| Mean (range) |
(2.7–153,580.0) (n=357)* |
(2.7–153,580.0) (n=85)* |
(9.7–69,019.0) (n=272)* |
|
| eGFR (mL/min/1.73 m2) |
36.2±19.8 |
36.3±27.4 |
36.2±16.6 |
0.963 |
| Hemoglobin |
11.5±2.1 (n=382)* |
11.8±2.1 |
11.5±2.1 (n=286)* |
0.183 |
| Comorbidities |
| Diabetes |
160 (41.7) |
37 (38.5) |
123 (42.7) |
0.550 |
| Hypertension |
238 (62.0) |
59 (61.5) |
179 (62.2) |
0.904 |
| Previous stroke |
53 (13.8) |
14 (14.6) |
39 (13.5) |
0.864 |
| AF |
250 (65.1) |
62 (64.6) |
188 (65.3) |
0.902 |
| HF therapy |
| LVEF <40% (n) |
260 |
63 |
197 |
|
| Medication |
| ACEi/ARB |
171 (65.8) |
37 (58.7) |
134 (68.0) |
0.222 |
| β-blocker |
229 (88.1) |
55 (87.3) |
174 (88.3) |
0.825 |
| Thiazide |
28 (10.8) |
8 (12.7) |
18 (9.1) |
0.428 |
| MRA |
135 (51.9) |
34 (54.0) |
101 (51.3) |
0.764 |
| Loop diuretic |
237 (91.2) |
56 (88.9) |
181 (91.9) |
0.453 |
| Tolvaptan |
183 (70.4) |
48 (76.2) |
135 (68.5) |
0.271 |
| NPPV |
| Prior ASV |
37 (14.2) |
34 (54.0) |
3 (1.5) |
<0.001 |
| Prior CPAP |
16 (6.2) |
3 (4.8) |
13 (6.6) |
0.768 |
| LVEF ≥40% (n) |
124 |
33 |
91 |
|
| Medication |
| ACEi/ARB |
70 (56.5) |
18 (54.5) |
52 (57.1) |
0.839 |
| β-blocker |
93 (75.0) |
26 (78.8) |
67 (73.6) |
0.644 |
| Thiazide |
11 (8.9) |
3 (9.1) |
8 (8.8) |
0.869 |
| MRA |
45 (36.3) |
11 (33.3) |
34 (37.4) |
0.881 |
| Loop diuretic |
112 (90.3) |
30 (90.9) |
82 (90.1) |
1.000 |
| Tolvaptan |
79 (63.7) |
25 (75.8) |
54 (59.3) |
0.138 |
| NPPV |
| Prior ASV |
16 (12.9) |
14 (42.4) |
2 (2.2) |
<0.001 |
| Prior CPAP |
8 (6.5) |
3 (9.1) |
5 (5.5) |
0.438 |
Unless indicated otherwise, data are presented as the mean±SD or n (%). *For variables with missing values, the number of patients evaluated is shown. Abbreviations as in Table 1.
Comparison of secondary endpoints in the PS-matched groups revealed a time to any cardiovascular event of 34.7 weeks (95% CI 19.7–66.7) in the ASV group and 58.7 weeks (95% CI 43.3–93.6) in the non-ASV group (P=0.008, log-rank test). The EQ-5D-5L indices at 1 year remained stable, with no significant change in walking ability (P=0.694), self-management ability (P=0.786), ordinal activity (P=0.707), or pain/discomfort (P=0.906). However, there was a significant deterioration of the anxiety/depression parameter in the ASV group (P=0.033; Supplementary Table 2). The clinical composite response improved in 7.7% of patients in the ASV group and in 16.3% of patients in the non-ASV group (P=0.081; Supplementary Table 2). A controlled NT-proBNP concentration at 1 year was achieved by 11.1% of patients in the ASV group and by 20.9% of patients in the non-ASV group (P=0.079; Supplementary Table 2).
Discussion
This multicenter prospective cohort study evaluated the impact of chronic home-based ASV therapy on all-cause mortality and the urgent rehospitalization rate in patients with chronic HF. Unlike in the SERVE-HF study,24 which suggested an increased all-cause mortality risk, the present study showed no significant effect on all-cause mortality. In terms of all-cause mortality and urgent rehospitalizations, ASV was associated with worse outcomes in patients with low NT-proBNP and comparable outcomes in those with high NT-proBNP, suggesting that ASV may be an option for selected patients with HF. In general, patients with more congestion (i.e., a high NT-proBNP concentration) are likely to derive more beneficial hemodynamic effects from ASV15 and to feel comfortable with it.
ASV was originally developed as an auxiliary device for managing SDB, including non-obstructive sleep apnea. Previous comparisons of ASV with nocturnal oxygen therapy or conventional NPPV in the treatment of SDB in patients with HF found that ASV effectively mitigated apneas, reduced the number of night-time awakenings, and improved cardiac function and QOL.28–31 However, SERVE-HF, which investigated the prognostic impact of ASV in patients with HF and SDB, found that although ASV was effective in suppressing SDB, it had no beneficial effect on HF or QOL and was associated with increased cardiovascular mortality.24
Although ASV was initially developed in Japan as a treatment for pulmonary congestion, numerous studies have demonstrated that it can improve LVEF, reduce the BNP concentration, and decrease HF-related mortality and hospitalizations.12–19 ASV has also been shown to suppress sympathetic nerve activity in patients with HF by activating parasympathetic stretch receptors in the lungs.19–23 Our multicenter retrospective SAVIOR-R study found that ASV improved LVEF and NYHA functional class in patients with chronic HF.11 We then investigated the effectiveness of chronic use of ASV further in the multicenter randomized controlled SAVIOR-C trial, in which improvements in cardiac events and symptoms (NYHA classification and physical ability) were observed in patients who received ASV but without a significant increase in LVEF.11 In view of these inconsistent findings, we initiated SAVIOR-L to prospectively assess the effects of chronic use of ASV in a real-world clinical setting.
Patients with recurrent hospitalizations for HF are typically categorized as Stage D in clinical guidelines. The present study represents the first focused investigation in these patients; its results were worse than anticipated and highlight the lack of effective therapies for these patients. We also evaluated the characteristics of patients with HF in whom chronic ASV was used; most were symptomatic and overtly congested despite treatment with loop diuretics and vasopressin V2
receptor antagonists, tended to be younger, and tended to have a history of more frequent hospitalizations than those in the non-ASV group. Interestingly, 50% of these patients had previously received home-based ASV therapy. Notably, chronic ASV was well tolerated with no reports of device failure.
We used PS matching in the present study to account for potential biases by adjusting for 9 specific covariates. However, a disparity in the NT-proBNP concentration remained between the study groups following PS matching. Therefore, we conducted post hoc analyses to account further for variation in the severity of HF. Nevertheless, there was a between-group difference in the hospitalization rate for HF at the start of the study, with the non-ASV group having a lower frequency of hospitalizations. This difference may have contributed substantially to the higher urgent rehospitalization rate in our ASV group. The Kaplan-Meier curves for all-cause mortality indicated comparable survival rates between the study groups.
Possible reasons for the association between chronic use of ASV and an increased urgent rehospitalization rate in patients with HF are as follows. First, ASV was used primarily to alleviate congestion. Default settings were used in most cases, with no device adjustments according to individual circumstances. This preference may have been influenced by previous reports demonstrating the impact of pulmonary pressure support levels on Frank-Starling curves.18 Furthermore, the possibility of a link between higher presets and increased cardiovascular mortality suggested by the SERVE-HF study may have led the attending physicians to exercise caution in setting more aggressive parameters. Second, patients in the ASV group already had a significantly greater number of previous hospitalizations for severe HF at baseline. Third, the higher all-cause urgent rehospitalization rate in the ASV group may reflect not only worsening HF, but also detection of relatively minor changes by the attending physicians, who may have been monitoring their patients on ASV more closely. These factors warrant further exploration. Nevertheless, the present study showed no association between the use of ASV and increased all-cause mortality.
Study Limitations
This study has some limitations. First, this study had an observational design and was not an interventional trial. Therefore, it was not possible to explore the possibility of a causal relationship between ASV and the poor prognosis in patients with recurrent hospitalizations for HF. Second, the PS-matching procedure had limited ability to adjust for background differences between the study groups. Aligning the background characteristics of such a critically ill cohort proved to be more challenging than anticipated. Further analyses will be conducted in the near future with adjustment for additional confounders. Third, the effects of chronic ASV were evaluated only in patients with advanced HF. Therefore, our findings cannot be extrapolated to patients with other stages of HF. Fourth, we have no data on Cheyne-Stokes respiration. Finally, confounding by indication bias could not be ruled out in either study group.
Conclusions
In this study, chronic use of ASV had no impact on all-cause mortality in patients with advanced HF. However, this treatment may increase the urgent rehospitalization rate in patients with low NT-proBNP. ASV may be an option for patients with a higher NT-proBNP concentration, provided that they demonstrate an ability to tolerate ASV.
Acknowledgments
The authors thank all the patients and investigators who contributed to this study. The authors are also grateful to Yuka Nakajima, Masafumi Mori, and other members of the Public Health Research Foundation for their administrative assistance and to the ASCA Corporation for medical writing support.
Sources of Funding
This study was supported by the Public Health Research Foundation, which receives funding from Teijin Pharma. Neither of these entities had a role in trial design, data analysis, data interpretation, or writing of the manuscript.
Disclosures
Y.F., H.T., and K.K. are members of Circulation Journal’s Editorial Team. T.T. has received lecture fees from Teijin Pharma. Y.F. and Y.K. have received research honoraria (lecture fees) from the Public Health Research Foundation and Teijin Pharma. S.M. has received lecture fees from Teijin Healthcare. H.T. has received remuneration (e.g., lecture fees) from Kowa, Teijin Pharma, Nippon Boehringer Ingelheim, Mitsubishi Tanabe Pharma, Pfizer Japan, Ono Pharmaceutical, Daiichi Sankyo, Novartis Pharma, Bayer Yakuhin, Otsuka Pharmaceutical, and AstraZeneca; manuscript fees from Nippon Rinsho; research funding from Mitsubishi Tanabe Pharma, Nippon Boehringer Ingelheim, IQVIA Services Japan, MEDINET, Medical Innovation Kyushu, Kowa, Daiichi Sankyo, Johnson & Johnson, and NEC; and scholarship funds or donations from Nippon Boehringer Ingelheim, St. Mary’s Hospital, Teijin Pharma, Daiichi Sankyo, and Mitsubishi Tanabe Pharma. T.K. is affiliated with a department endowed by Philips, Fukuda Denshi, and ResMed. The other authors report no conflicts of interest.
IRB Information
This study was approved by the Institutional Review Board of Public Health Research Foundation (Reference no. 18G0001).
Data Availability
The deidentified participant data and further details of the study protocol will be shared upon request for up to 24 months after publication of this article. Researchers who request this information should include a methodologically sound proposal on how the data will be used. The proposal will be reviewed by the working committee, and a data access agreement will need to be signed. The data will be shared as Excel files via email. Proposals should be directed to ykihara@kcho.jp.
Supplementary Files
Please find supplementary file(s);
https://doi.org/10.1253/circj.CJ-23-0827
References
- 1.
Tsutsui H, Ide T, Ito H, Kihara Y, Kinugawa K, Kinugawa S, et al. JCS/JHFS 2021 guideline focused update on diagnosis and treatment of acute and chronic heart failure. J Card Fail 2021; 27: 1404–1444, doi:10.1016/j.cardfail.2021.04.023.
- 2.
Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure: A report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2022; 145: e895–e1032, doi:10.1161/CIR.0000000000001063.
- 3.
McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Bohm M, et al. 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur J Heart Fail 2022; 24: 4–131, doi:10.1002/ejhf.2333.
- 4.
Okura Y, Ramadan MM, Ohno Y, Mitsuma W, Tanaka K, Ito M, et al. Impending epidemic: Future projection of heart failure in Japan to the year 2055. Circ J 2008; 72: 489–491, doi:10.1253/circj.72.489.
- 5.
Ushigome R, Sakata Y, Nochioka K, Miyata S, Miura M, Tadaki S, et al. Temporal trends in clinical characteristics, management and prognosis of patients with symptomatic heart failure in Japan: Report from the CHART studies. Circ J 2015; 79: 2396–2407, doi:10.1253/circj.CJ-15-0514.
- 6.
Yukino M, Nagatomo Y, Goda A, Kohno T, Takei M, Nishihata Y, et al. Association of non-invasive positive pressure ventilation with short-term clinical outcomes in patients hospitalized for acute decompensated heart failure. J Clin Med 2021; 10: 5092, doi:10.3390/jcm10215092.
- 7.
Naughton MT, Rahman MA, Hara K, Floras JS, Bradley TD. Effect of continuous positive airway pressure on intrathoracic and left ventricular transmural pressures in patients with congestive heart failure. Circulation 1995; 91: 1725–1731, doi:10.1161/01.cir.91.6.1725.
- 8.
Mehta S, Hill NS. Noninvasive ventilation. Am J Respir Crit Care Med 2001; 163: 540–577, doi:10.1164/ajrccm.163.2.9906116.
- 9.
Takano T, Endo T, Tanaka K, Seino Y, Nitta T, Matsuyama Y, et al. Effects of positive end-expiratory pressure ventilation and extracorporeal ultrafiltration method in patients with refractory heart failure. Jpn Circ J 1986; 50: 359–367, doi:10.1253/jcj.50.359.
- 10.
Gray A, Goodacre S, Newby DE, Masson M, Sampson F, Nicholl J, et al. Noninvasive ventilation in acute cardiogenic pulmonary edema. N Engl J Med 2008; 359: 142–151, doi:10.1056/NEJMoa0707992.
- 11.
Momomura S, Seino Y, Kihara Y, Adachi H, Yasumura Y, Yokoyama H. Adaptive servo-ventilation therapy using an innovative ventilator for patients with chronic heart failure: A real-world, multicenter, retrospective, observational study (SAVIOR-R). Heart Vessels 2015; 30: 805–817, doi:10.1007/s00380-014-0558-8.
- 12.
Takama N, Kurabayashi M. Effectiveness of adaptive servo-ventilation for treating heart failure regardless of the severity of sleep-disordered breathing. Circ J 2011; 75: 1164–1169, doi:10.1253/circj.CJ-10-0831.
- 13.
Koyama T, Watanabe H, Igarashi G, Terada S, Makabe S, Ito H. Short-term prognosis of adaptive servo-ventilation therapy in patients with heart failure. Circ J 2011; 75: 710–712, doi:10.1253/circj.CJ-10-0956.
- 14.
Haruki N, Takeuchi M, Kaku K, Yoshitani H, Kuwaki H, Tamura M, et al. Comparison of acute and chronic impact of adaptive servo-ventilation on left chamber geometry and function in patients with chronic heart failure. Eur J Heart Fail 2011; 13: 1140–1146, doi:10.1093/eurjhf/hfr103.
- 15.
Yamada S, Sakakibara M, Yokota T, Kamiya K, Asakawa N, Iwano H, et al. Acute hemodynamic effects of adaptive servo-ventilation in patients with heart failure. Circ J 2013; 77: 1214–1220, doi:10.1253/circj.CJ-12-1088.
- 16.
Takama N, Kurabayashi M. Safety and efficacy of adaptive servo-ventilation in patients with severe systolic heart failure. J Cardiol 2014; 63: 302–307, doi:10.1016/j.jjcc.2013.09.008.
- 17.
Hiasa G, Okayama H, Hosokawa S, Kosaki T, Kawamura G, Shigematsu T, et al. Beneficial effects of adaptive servo-ventilation therapy on readmission and medical costs in patients with chronic heart failure. Heart Vessels 2018; 33: 859–865, doi:10.1007/s00380-018-1124-6.
- 18.
Imamura T, Kinugawa K. Right ventricular end-diastolic pressure is a key to the changes in cardiac output during adaptive servo-ventilation support in patients with heart failure. Int Heart J 2017; 58: 536–543, doi:10.1536/ihj.16-489.
- 19.
Koyama T, Watanabe H, Tamura Y, Oguma Y, Kosaka T, Ito H. Adaptive servo-ventilation therapy improves cardiac sympathetic nerve activity in patients with heart failure. Eur J Heart Fail 2013; 15: 902–909, doi:10.1093/eurjhf/hft049.
- 20.
Harada D, Joho S, Oda Y, Hirai T, Asanoi H, Inoue H. Short term effect of adaptive servo-ventilation on muscle sympathetic nerve activity in patients with heart failure. Auton Neurosci 2011; 161: 95–102, doi:10.1016/j.autneu.2010.12.001.
- 21.
Joho S, Oda Y, Ushijima R, Hirai T, Inoue H. Effect of adaptive servoventilation on muscle sympathetic nerve activity in patients with chronic heart failure and central sleep apnea. J Card Fail 2012; 18: 769–775, doi:10.1016/j.cardfail.2012.08.360.
- 22.
Ushijima R, Joho S, Akabane T, Oda Y, Inoue H. Differing effects of adaptive servoventilation and continuous positive airway pressure on muscle sympathetic nerve activity in patients with heart failure. Circ J 2014; 78: 1387–1395, doi:10.1253/circj.CJ-13-1468.
- 23.
Tamura Y, Koyama T, Watanabe H, Hosoya T, Ito H. Beneficial effects of adaptive servo-ventilation therapy on albuminuria in patients with heart failure. J Cardiol 2015; 65: 412–417, doi:10.1016/j.jjcc.2014.07.008.
- 24.
Cowie MR, Woehrle H, Wegscheider K, Angermann C, d’Ortho MP, Erdmann E, et al. Adaptive servo-ventilation for central sleep apnea in systolic heart failure. N Engl J Med 2015; 373: 1095–1105, doi:10.1056/NEJMoa1506459.
- 25.
Kihara Y, Seino Y, Momomura S; SAVIOR-C Investigators. Adaptive servo-ventilation for central sleep apnea in heart failure. N Engl J Med 2016; 374: 687–688.
- 26.
Tsutsui H, Isobe M, Ito H, Ito H, Okumura K, Ono M, et al. JCS 2017/JHFS 2017 guideline on diagnosis and treatment of acute and chronic heart failure: Digest version. Circ J 2019; 83: 2084–2184, doi:10.1253/circj.CJ-19-0342.
- 27.
Ishihara S, Hiramitsu S, Kanaoka K, Taki M, Nakagawa H, Ueda T, et al. New conversion formula between B-type natriuretic peptide and N-terminal-pro-B-type natriuretic peptide: Analysis from a multicenter study. Circ J 2022; 86: 2010–2018, doi:10.1253/circj.CJ-22-0032.
- 28.
Teschler H, Döhring J, Wang YM, Berthon-Jones M. Adaptive pressure support servo-ventilation: A novel treatment for Cheyne-Stokes respiration in heart failure. Am J Respir Crit Care Med 2001; 164: 614–619, doi:10.1164/ajrccm.164.4.9908114.
- 29.
Morgenthaler TI, Gay PC, Gordon N, Brown LK. Adaptive servoventilation versus noninvasive positive pressure ventilation for central, mixed, and complex sleep apnea syndromes. Sleep 2007; 30: 468–475, doi:10.1093/sleep/30.4.468.
- 30.
Pepperell JC, Maskell NA, Jones DR, Langford-Wiley BA, Crosthwaite N, Stradling JR, et al. A randomized controlled trial of adaptive ventilation for Cheyne-Stokes breathing in heart failure. Am J Respir Crit Care Med 2003; 168: 1109–1114, doi:10.1164/rccm.200212-1476OC.
- 31.
Philippe C, Stoica-Herman M, Drouot X, Raffestin B, Escourrou P, Hittinger L, et al. Compliance with and effectiveness of adaptive servoventilation versus continuous positive airway pressure in the treatment of Cheyne-Stokes respiration in heart failure over a six month period. Heart 2006; 92: 337–342, doi:10.1136/hrt.2005.060038.
