Background
Left atrial (LA) dysfunction has been related to symptom onset in patients with heart failure (HF). However, the potential prognostic role of LA function has been scarcely studied in outpatients with new-onset HF symptoms.
Methods
Consecutive outpatients with suspected HF onset evaluated at a one-stop clinic were screened. HF diagnosis was performed according to current guidelines. LA function was analyzed in patients in sinus rhythm by speckle-tracking echocardiography, determining LA peak strain rate after atrial contraction (LASRa) as a surrogate of atrial contractile function. Yearly prospective follow-up was conducted to report cardiovascular hospital admission or death. Patients without HF in sinus rhythm were followed as a control group. Survival curves were estimated using the Kaplan-Meier method.
Results
One hundred fifty-four outpatients were included (mean age, 74 ± 10 years; 67% women) with a median follow-up duration of 44.4 months (interquartile range, 31–58 months). Final diagnosis was 29.9% non-HF and 70.1% HF. More than two in five patients with HF (44.4%) had AF ( n = 48), and 55.6% ( n = 60) were in sinus rhythm. The latter were divided according to LASRa tertile: highest, −1.93 ± 0.39 sec −1 ; middle, −1.08 ± 0.21 sec −1 ; and lowest, −0.47 ± 0.18 sec −1 . At the end of follow-up, patients with atrial fibrillation had a low event-free survival rate (56.3%), similar to those in the lower LASRa tertile (55.0%). The non-HF group had the best prognosis, and the higher and middle LASRa tertiles had intermediate prognoses (event-free survival, 85%, 75%, and 70%, respectively).
Conclusions
The study of contractile LA function in outpatients with new-onset HF provides prognostic stratification. The early identification of patients at higher risk on the basis of their atrial function would allow focusing on them independently of their final diagnoses.
Heart failure (HF) with reduced ejection fraction (HFREF) can be easily diagnosed using echocardiography by a reduced left ventricular (LV) ejection fraction. HF with preserved ejection fraction (HFPEF) is more difficult to diagnose, although it is the most prevalent type of HF in outpatients and has high morbidity and mortality. The diagnosis is based on the Paulus algorithm, which includes B-type natriuretic peptide (BNP) determination and different echocardiographic measures. It is difficult to use in normal clinical practice, especially in outpatients. BNP has also shown utility for prognosis in patients with HF.
Left atrial (LA) function can be easily studied using speckle-tracking strain echocardiography. LA function is divided into three phases: reservoir (filling during ventricular systole), conduit (passive emptying during early ventricular diastole), and active LA contraction (late ventricular diastole). In a previous work with an earlier inclusion period of patients at our HF clinic, we found that impaired atrial strain was related to symptom onset in patients with HF. In this group of outpatients with new-onset HF, LA strain was similarly reduced in patients with HFREF and those with HFPEF, but whereas in patients with HFREF, LV longitudinal strain was reduced, LV longitudinal strain showed no differences between patients with HFPEF and a non-HF group, suggesting earlier involvement of the LA in patients with HF.
LA strain was recently related to cardiovascular outcomes and new-onset atrial fibrillation (AF) in the general population. LV strain was also related to cardiovascular prognosis in patients with suspected HF. The prognostic value of LA function has also been explored using other techniques. LA function measured using magnetic resonance as LA ejection fraction was related to clinical outcomes in patients with HF. The assessment of LA work by computation as [LA stroke volume × blood density × (transmitral Doppler peak atrial velocity) 2 ] showed incremental prognostic value over LA size in patients with chronic HF to predict death and HF hospitalization. Finally, the measurement of an LA function index as (LA ejection fraction × velocity-time integral of the LV outflow tract × LA maximal volume) in patients with first hospital admissions for HFREF significantly predicted adverse events in the first 6 months of follow-up. However, the prognostic significance of impaired LA strain in patients with HF, particularly in the outpatient setting, is unknown. Additionally, more precise risk stratification of outpatients with potential symptoms of early HF that would allow us to focus on higher risk patients is needed. We hypothesized that single study of LA function could stratify the cardiovascular risk of outpatients with new-onset HF. Therefore, we aimed to evaluate the usefulness of the analysis of atrial contractile function to predict cardiovascular outcomes in outpatients with HF onset.
Methods
Study Design and Ethics
This was an observational study with prospective screening of outpatients referred for diagnostic workup to our HF clinic between March 2009 and March 2014. Longitudinal follow-up was conducted to report death or hospital admission for a cardiovascular reason. The performance of the one-stop HF clinic was previously described elsewhere. The present study represents the later phase of follow-up of outpatients included in our cohort of patients with new-onset HF; some of the patients included in the present study were also part of previous studies (which involved smaller numbers of patients and shorter duration of follow-up). Despite using similar and overlapped populations, the objectives of the previous studies were distinct and also complementary to those of the present study, as follows: in one previous study, the objective was to analyze the role of LA function in the differential diagnosis of outpatients with new-onset symptoms of HF, while in the other study, we examined the relation of blood biomarkers with prognosis (also including visits to the emergency department for cardiovascular reasons as an end point).
The investigation conformed with the principles outlined in the Declaration of Helsinki. The study protocol was approved by the ethics committee of our institution, and all participants provided written informed consent.
Patients
The cohort of consecutive patients visiting the HF clinic for suspected new onset of HF formed the population studied. At inclusion, diagnosis was performed according to current recommendations as non-HF, HFPEF, or HFREF. The initial visit included physical examination, electrocardiography, chest radiography, blood test with BNP measurement, and transthoracic echocardiography. Exclusion criteria were age < 18 years, life expectancy < 1 year, and/or inability to perform the diagnostic circuit as previously described. Patients with final diagnosis of non-HF and sinus rhythm were included as a control group.
Echocardiographic Acquisition and Analysis
A two-dimensional echocardiographic study with conventional Doppler and tissue Doppler was performed using a commercially available system (Vivid 7; GE Healthcare, Milwaukee, WI). LV and LA dimensions were determined according to current recommendations and indexed to body surface area (calculated using the Du Bois method). LA deformation was analyzed in patients in sinus rhythm from two-dimensional echocardiography using dedicated software (two-dimensional strain, EchoPAC version 112; GE Healthcare, Milwaukee, WI). The frame rate was set between 60 and 80 frames/sec to ensure adequate speckle-tracking. The onset to analyze strain was determined by the onset of the P wave on the electrocardiogram. LA longitudinal deformation was quantified and averaged from six segments of the left atrium from the apical four-chamber view. LA peak systolic strain rate (LASRs) as a surrogate of LA reservoir function, LA peak strain rate after atrial contraction (LASRa) as a surrogate of LA contractile function, and LA peak strain rate during early ventricular diastole (LASRe) as a surrogate of conduit function were determined ( Figure 1 ). Reproducibility analyses for LA strain measurements were performed by two investigators in 10 consecutive patients in sinus rhythm after all echocardiographic measurements had been completed. The new measurements were performed blinded to the initial results and the results of the other investigator. The measurements were performed in the first suitable video clip, but the electrocardiographic cycle to analyze was selected by the investigator at the time of the new measurement. Atrial strain in patients with AF was not measured, because of its high beat-to-beat variability. Global longitudinal LV strain was also quantified and averaged from six myocardial LV segments in the apical four-chamber view.
Follow-Up
Cardiovascular events were prospectively reported during follow-up. A telephone interview and a medical history review of the centralized digital medical records of our health network referral area were conducted yearly by the research team. The duration of follow-up was the interval between the date of inclusion (initial visit to the outpatient HF clinic) and the date of last contact or death. A composite end point was defined to evaluate cardiovascular outcomes, including all-cause death or cardiovascular hospitalization (HF, acute coronary syndrome, arrhythmia, sudden death).
Statistical Analysis
Data are expressed as mean ± SD, frequency distribution, or proportions as appropriate. Normality of distribution of quantitative variables was assessed using the Kolmogorov-Smirnov test. A descriptive and comparative analysis was performed between the different diagnostic groups. The χ 2 test or Fisher exact test was used to compare categorical variables, and Student’s t test was used for independent samples of quantitative variables. Analysis of variance and Bonferroni statistical tests were used to compare quantitative variables between more than two groups. Receiver operating characteristic curves were assessed to identify correlations of echocardiographic parameters with cardiovascular hospital admission or death. Survival curves for patient groups were estimated using the Kaplan-Meier product-limit estimator, and these were compared using the log-rank test. The Pearson test was used to correlate quantitative variables. Intraobserver and interobserver intraclass correlations for LA strain analysis were performed using Cronbach’s α method. P values < .05 (two sided) were considered to indicate statistical significance. Data were processed using SPSS version 19 (IBM, Armonk, NY).
Results
Demographics and Clinical Data
One hundred fifty-four patients were included, with a median follow-up duration of 44.4 months (interquartile range, 31–58 months). Figure 2 shows the distribution of included patients in the different groups. The mean age was 74 ± 1 years, 67% were women, and the most prevalent cardiovascular risk factor was systemic arterial hypertension (77%). The final diagnosis was HF in 70.13% patients ( n = 108) (68.5% with HFPEF and 31.5% with HFREF). The New York Heart Association functional class distribution of patients with HF at the moment of inclusion was 1.8% in class I, 61.1% in class II, and 37.1% in class III. Baseline and echocardiographic characteristics by diagnostic groups are shown in Table 1 . AF was present at inclusion in 44.4% of patients with HF. LV ejection fraction and global longitudinal LV strain showed no differences between the non-HF and HFPEF groups and were significantly decreased in the HFREF group. LA volume was increased in both HF groups compared with the non-HF group. LA strain rate was measured in patients in sinus rhythm ( n = 116) and was similarly decreased in both HF groups compared with the control group.
| Baseline characteristics | HFPEF ( n = 74) | HFREF ( n = 34) | Non-HF ( n = 46) | P value | ||
|---|---|---|---|---|---|---|
| Non-HF vs HFPEF | Non-HF vs HFREF | HFPEF vs HFREF | ||||
| Age (y) | 76.6 ± 7.2 | 74.2 ± 12.3 | 71.3 ± 11.4 | 0.013 | 0.560 | 0.731 |
| Women | 74.3 (55) | 38.2 (13) | 76.1 (35) | 0.169 | 0.001 | <0.001 |
| Hypertension | 86.5 (64) | 73.5 (25) | 63 (29) | 0.009 | 0.756 | 0.401 |
| Diabetes | 31.1 (23) | 41.2 (14) | 23.9 (11) | 1.000 | 0.305 | 0.096 |
| Tobacco exposure | 28.4 (21) | 55.9 (19) | 32.6 (15) | 1.000 | 0.092 | 0.016 |
| AF | 39.2 (29) | 55.9 (19) | 0 (0) | <0.001 | <0.001 | 0.162 |
| HR (beats/min) | 69.6 ± 12.6 | 81.3 ± 16.4 | 74.9 ± 10.2 | 0.099 | 0.088 | <0.001 |
| BNP (ng/mL) | 159.4 ± 121.1 | 378.0 ± 558.7 | 35.8 ± 22.5 | 0.053 | <0.001 | 0.001 |
| LVDV (mL/m 2 ) | 56.0 ± 19.0 | 98.6 ± 35.9 | 51.1 ± 23.0 | 0.969 | <0.001 | <0.001 |
| LVEF (%) | 59.8 ± 5.2 | 35.1 ± 9.7 | 60.8 ± 3.7 | 1.000 | <0.001 | <0.001 |
| LV global strain (%) | −16.8 ± 3.62 | −9.9 ± 4.5 | −17.3 ± 3.9 | 1.000 | <0.001 | <0.001 |
| LA volume (mL/m 2 ) | 58.6 ± 23.3 | 58.8 ± 20.5 | 14.8 ± 2.2 | <0.001 | <0.001 | 1.000 |
| LAS (%) | 18.37 ± 6.85 | 17.31 ± 10.40 | 25.41 ± 6.09 | <0.001 | 0.001 | 1.000 |
| LASRs (sec −1 ) | 0.94 ± 0.35 | 0.72 ± 0.44 | 1.43 ± 0.45 | <0.001 | <0.001 | 0.227 |
| LASRa (sec −1 ) | −1.18 ± 0.68 | −1.10 ± 0.60 | −1.99 ± 0.58 | <0.001 | <0.001 | 1.000 |
| LASRe (sec −1 ) | −0.72 ± 0.41 | −0.67 ± 0.32 | −0.90 ± 0.49 | 0.147 | 0.205 | 1.000 |
| Diastolic dysfunction grade | 1.6 ± 0.6 | 2.1 ± 0.8 | 0.8 ± 0.4 | <0.001 | <0.001 | 0.029 |
Follow-Up
All patients included were followed for a median duration of 44.4 months (interquartile range, 31–58 months). During follow-up, 48 patients had at least one event (death or hospital admission for cardiovascular reason). There were 25 deaths, 20 in the HF group (12 in the AF group, eight in the sinus rhythm group) and five in the control group. Thirty-six patients had at least one hospital admission for cardiovascular reasons (33 in the HF group and three in the control group). The most common cause of hospital admission was HF (61.1% [ n = 22]); other causes were AF ( n = 8), AF and stroke ( n = 2), acute coronary syndrome with HF ( n = 1), sudden death ( n = 1), and cardiogenic syncope ( n = 1). Time from inclusion to the first event was used to construct Kaplan-Meier survival curves.
Receiver operating characteristic curves for the different prognostic echocardiographic parameters, including LA indexed volume, LV global strain, and LA strain and strain rate, were constructed to predict events ( Figure 3 ). The areas under the curves were as follows: LASRa, 0.739 (95% CI, 0.630–0.848; P = .001); LASRs, 0.702 (95% CI, 0.587–0.817; P = .004); LASRe, 0.682 (95% CI, 0.562–0.802; P = .010), LA strain, 0.715 (95% CI, 0.591–0.839; P = .002); LA volume index, 0.678 (95% CI, 0.555–0.801; P = .011); and LV longitudinal strain, 0.635 (95% CI, 0.503–0.767; P = .056).
Because of its larger area under the curve, LASRa was the selected LA parameter to evaluate cardiovascular prognosis (cutoff value of LASRa for event prediction, −1.400 sec −1 [63.6% sensitivity and 65.3% specificity]). Patients with HF in sinus rhythm were divided into tertiles according to LASRa. Patients with HF with AF were followed as a separate group ( Figure 2 ).
Table 2 shows the characteristics of the studied population in relationship to the LA function groups. Age and sex were similar between groups, while cardiovascular risk factors were less prevalent in the non-HF group. As anticipated, LV diastolic dysfunction was more impaired in the AF and low-LASRa groups. In patients in sinus rhythm, LASRa had moderate, though significant, correlations with BNP (0.315, P = .001), LA indexed volume (0.580, P < .001), LV indexed volume (0.371, P < .001), and LV longitudinal strain (0.334, P = .001).
| Baseline characteristics | Control group ( n = 46) | High tertile of LASRa ( n = 20) | Middle tertile of LASRa ( n = 20) | Low tertile of LASRa ( n = 20) | AF ( n = 48) | Global P value |
|---|---|---|---|---|---|---|
| Age (y) | 71.3 ± 11.4 | 76.9 ± 10.3 | 73.9 ± 10.6 | 73.9 ± 7.2 | 77.0 ± 8.7 | 0.057 |
| Women | 76.1 (35) | 70.0 (14) | 70.0 (14) | 60.0 (12) | 58.3 (28) | 0.421 |
| Hypertension | 63.0 (29) | 90.0 (18) | 90.0 (18) | 85.0 (17) | 75.0 (36) | 0.047 |
| Diabetes | 23.9 (11) | 30.0 (6) | 30.0 (6) | 55.0 (11) | 29.2 (14) | 0.164 |
| Smokers | 32.6 (15) | 30.0 (6) | 25.0 (5) | 40.0 (8) | 43.8 (21) | 0.568 |
| HR (beats/min) | 74.9 ± 10.1 | 72.3 ± 13.9 | 72.9 ± 11.7 | 65.5 ± 11.4 | 77.2 ± 16.6 | 0.025 |
| BNP (ng/mL) | 35.8 ± 22.5 | 92.5 ± 73.4 | 174.2 ± 157.0 | 388.1 ± 707.3 | 240.7 ± 175.9 | <0.001 |
| HFREF | — | 20.0 (4) | 25.0 (5) | 30.0 (6) | 39.6 (19) | 0.386 |
| LVDV (mL/m 2 ) | 51.1 ± 23.0 | 65.2 ± 24.6 | 77.96 ± 30.2 | 87.73 ± 48.9 | 61.2 ± 24.9 | <0.001 |
| LVEF (%) | 60.8 ± 3.7 | 56.1 ± 11.0 | 52.8 ± 14.3 | 48.5 ± 17.2 | 51.4 ± 12.0 | <0.001 |
| LV global strain (%) | −17.3 ± 3.9 | −17.5 ± 4.1 | −14.4 ± 5.5 | −12.4 ± 4.3 | −13.8 ± 5.2 | <0.001 |
| LA volume (mL/m 2 ) | 31.6 ± 14.8 | 46.7 ± 15.2 | 54.5 ± 13.4 | 62.4 ± 21.3 | 63.8 ± 26.3 | <0.001 |
| LAS (%) | 25.41 ± 6.09 | 22.35 ± 8.59 | 17.39 ± 3.64 | 14.59 ± 8.34 | — | <0.001 |
| LASRs (sec −1 ) | 1.42 ± 0.45 | 1.20 ± 0.39 | 0.81 ± 0.14 | 0.65 ± 0.34 | — | <0.001 |
| LASRa (sec −1 ) | −1.99 ± 0.58 | −1.93 ± 0.39 | −1.08 ± 0.21 | −0.47 ± 0.18 | — | <0.001 |
| LASRe (sec −1 ) | −0.91 ± 0.49 | −0.86 ± 0.49 | −0.61 ± 0.30 | −0.67 ± 0.33 | — | 0.035 |
| Diastolic dysfunction grade | 0.8 ± 0.4 | 1.3 ± 0.6 | 1.5 ± 0.7 | 2.1 ± 0.6 | 1.9 ± 0.7 | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
