Abstract
Introduction
Left ventricular diastolic dysfunction (LVDD) frequently occurs in haemodialysis patients and is associated with adverse outcomes. Lung ultrasound (LUS) has been recently proposed for the quantification of extravascular lung water through assessment of B-lines. LUS findings and their relationship with LVDD in clinically euvolemic haemodialysis patients were investigated in this study.
Methods
Echocardiography and LUS examinations were performed on each patient. Multivariate linear regression and forward stepwise logistic regression were performed to determine the relationship between B-lines and LVDD. A receiver operating characteristic (ROC) curve with area under the curve (AUC) was calculated to determine the accuracy of B-lines for evaluating LVDD.
Results
A total of 119 patients were enrolled. The number of B-lines was statistically related to echocardiographic parameters (LAVI, LVEDVI, E/A, and E/e′) of diastolic function, while the relationship between B-lines and LVEF disappeared after adjusting for potential confounding factors. Additionally, compared with the mild B-line group (B-lines: <14), the moderate (B-lines: 14–30) and severe B-line groups (B-lines: >30) were associated with an increased risk of LVDD (OR 24.344, 95% CI 4.854–122.084, p < 0.001, and OR 94.552, 95% CI 9.617–929.022, p < 0.001, respectively). Furthermore, the AUC of the ROC curve for B-lines predicting LVDD was 0.845, and the cut-off of B-lines was 14.5 (sensitivity 64.91%, specificity 93.55%).
Conclusion
LUS B-lines were closely associated with left ventricular diastolic function in clinically euvolemic haemodialysis patients. Moreover, our findings suggested a B-line ≥14.5 as a reliable cut-off value for identifying patients with LVDD. LUS B-lines may be used as a novel indicator for evaluating LVDD.
Keywords: B-lines, Haemodialysis, Lung ultrasound, Left ventricular diastolic function
Introduction
Cardiovascular disease (CVD) is the main cause of mortality in haemodialysis patients, accounting for approximately 50% of all deaths in this population based on data from European countries and the USA [1, 2]. The reported prevalence of left ventricular diastolic dysfunction (LVDD) in haemodialysis patients varies from 25% to 87% [3, 4]. LVDD has been proven to predict a worse cardiovascular prognosis in haemodialysis patients [5, 6]. Therefore, early and simple detection of LVDD is crucial for end-stage renal disease (ESRD) patients on dialysis.
Ideally, measurement of left ventricle (LV) diastolic function should be obtained invasively by the gold standard method, cardiac catheterization, which has significant risk and costs; therefore, it cannot be directly applied in routine clinical practice [7]. Echocardiography is frequently used for noninvasive assessment of diastolic function [7]. However, algorithms of LVDD, based on the estimation of multiple parameters by tissue Doppler imaging, are quite complex and time-consuming [8].
In recent years, lung ultrasound (LUS) has been increasingly used to objectively quantify extravascular lung water (ELW) through the analysis of B-line artifacts [9], defined as discrete laser-like vertical hyperechoic reverberation artifacts that arise from the pleural line and extend to the bottom of the screen without fading and while moving in tandem with lung sliding [10]. This bedside technique is faster to perform, is less expensive, and has lower technical requirements than a full echocardiography examination. ELW, determined by the filling pressure of the LV, is common in haemodialysis patients [11–13] and is usually asymptomatic. It has been suggested that the effect of LV diastolic function on ELW is greater than that of systolic function [14]. Furthermore, previous studies reported that B-lines were associated with several echocardiography parameters, including left atrial volume, pulmonary artery pressure, peak early (E)/early peak diastolic annular velocity (e′) ratio, and LV ejection fraction (LVEF), in haemodialysis patients [11, 13, 14]. However, data regarding the association between B-lines and LVDD assessed comprehensively by the recommendations of echocardiographic algorithms, instead of a single indicator such as E/peak late (A) or E/e′, in clinically euvolemic haemodialysis patients are still lacking. Our aim in this study, therefore, was to test the hypothesis that B-lines are associated with LVDD, evaluated by the 2016 American Society of Echocardiography/European Association of Cardiovascular Imaging (ASE/EACVI) recommendations [15], in clinically euvolemic haemodialysis patients.
Materials and Methods
Study Population
This observational study enrolled patients from the haemodialysis centre in Tongren Hospital, Shanghai Jiao Tong University School of Medicine, between August 2021 and September 2022. Patients were eligible for study inclusion if they were adults (>18 years) with ESRD who were treated with thrice-weekly maintenance haemodialysis for ≥3 months and were euvolemic under standard clinical criteria. Exclusion criteria included chronic obstructive pulmonary disease or pulmonary interstitial disease, dry weight modification during 1 month before study enrolment, inadequate lung scanning and echocardiographic study, significant arrhythmias, severe mitral valve disease, underlying malignancy, current treatment for infection, history of drug or known severe mental disorder, and refusal to give consent. Our study was approved by the Ethics Committee of the Tongren Hospital Affiliated to Shanghai Jiao Tong University (No. 2021-008-01).
Sample Size
The sample size was calculated based on the formula for cross-sectional studies (n = Z21−α/2 × P × [1−P]/d2). According to the reported prevalence of LVDD in haemodialysis patients [3, 4], we set P at 87%, α at 5%, and d at 10%, meaning that at least 44 patients were needed to test our hypothesis.
Clinical and Laboratory Measurements
Baseline evaluation consisted of demographic characteristics, medical history, drug treatment, and a detailed physical examination. Evaluation was performed during the first or the second interdialytic interval of the week.
At study entry, blood was collected after an overnight fast on the same day of LUS and echocardiogram examinations. Assays for serum haemoglobin, calcium, phosphorus, intact parathyroid hormone, albumin, C-reactive protein, and B-type natriuretic peptide (BNP) were performed at the central laboratory of Tongren Hospital.
Echocardiography and LUS
According to a previous study [16], both echocardiography and LUS evaluations were performed on a short interdialytic day, 24 h after the scheduled starting time of the previous dialysis. The operators were blinded to the baseline evaluation.
Echocardiograms were performed by a standard ultrasound device (SIMENS ACUSON Sequoia Model 512 with a 3.5 MHz transducer). Two-dimensional measurements and tissue Doppler imaging were performed on all patients according to recommendations [17]. Measurements were averaged over 3 cardiac cycles. The left atrial volume index (LAVI), LV end-diastolic volume index (LVEDVI), LV end-systolic volume index, LVEF, LV mass index, E, A transmitral flow velocities, ratio of early/late peak velocities (E/A), e′, and ratio of the early transmitral flow velocity to early diastolic mitral annular velocity (E/e′) were recorded and calculated. The above indices were normalized for body surface area where appropriate. LVDD was defined according to the 2016 recommendations of the ASE/EACVI [15]. In patients with preserved LVEF, LVDD is diagnosed if more than two of the following parameters meet the pathologic threshold: average lateral and septal E/e′ >14, lateral e′ <10 cm/s or septal e′ <7 cm/s, LAVI >34 mL/m2, and tricuspid regurgitation velocity >2.8 m/s.
The examinations of B-lines were performed with patients in the supine or near-supine position. According to a more quantitative method used in nephrology [18], both sides of the anterior and lateral chest are scanned from the second to the fourth (on the left side) or fifth (on the right side), along the parasternal, midclavicular, anterior axillary, and midaxillary lines, for a total of 28 scanning sites. B-line number is recorded as a value from 0 to 10 in each intercostal space. The sum of B-lines produces a score that quantifies the degree of ELW. On the basis of this score [11], we grouped the patients into three categories of increasingly severe ELW (none or mild: <14 B-lines, moderate: 14–30 B-lines, and severe: >30 B-lines).
Statistical Analysis
Continuous data with normally distributed parameters are presented as the means ± standard deviations, and categorical variables are presented as percentages. In the case of nonparametric data distribution, medians with interquartile ranges are presented. Comparison of data with a normal distribution was made using an independent sample t test or one-way ANOVA. We used Bonferroni correction if the variance was homogeneous, and we used Tamhane’s T2 test if not. Nonnormally distributed data were compared using the Kruskal-Wallis test. Categorical data were compared using the χ2 test.
Spearman’s correlation and stepwise multivariate linear regression analyses were used to analyse the association of B-lines with echocardiographic parameters. B-lines and BNP were log-transformed to approximate normal distribution. Forward stepwise (likelihood ratio) logistic regression analyses were performed to investigate the correlation between B-lines and LVDD, with an odds ratio (OR) and 95% confidence interval (95% CI). A receiver operating characteristic (ROC) curve was created, and areas under the curve were calculated for the ability of B-lines to identify patients with LVDD. Results with p values <0.05 were considered statistically significant. Statistical analysis was performed using SPSS 26.0.
Results
Baseline Characteristics by the B-Line Category
Figure 1 shows the flowchart of the study. We finally included 119 participants with a mean age of 62.6 ± 13.1 years and 81 males (68.1%). Baseline characteristics by B-line category are summarized in Table 1. The proportion of CVD and diabetes increased in the B3 (B-lines >30) group compared with the B2 (14≤ B-lines ≤30) group. Participants in the B3 group had lower levels of albumin, calcium, and LVEF; higher levels of C-reactive protein, BNP, LAVI, LVEDVI, LV end-systolic volume index, LV mass index, e′, E/A, E/e′, and pulmonary pressure; and a higher proportion of LVDD than those in the B1 group (all p < 0.05; Table 1). There were no significant differences in the use of various antihypertensive drugs or aetiologies of chronic kidney disease among these groups.
Fig. 1.
Flow diagram of the study on the association of LUS B-lines with left ventricular diastolic function in clinically euvolemic haemodialysis patients.
Table 1.
Characteristics of the study population according to the grading of LUS B-lines
| Variables | All (n = 119) | LUS B-line scores | p value | ||
|---|---|---|---|---|---|
| <14 | 14–30 | >30 | |||
| B1 (n = 77) | B2 (n = 18) | B3 (n = 24) | |||
| Male gender, n (%) | 81 (68.1) | 48 (62.3) | 16 (88.9) | 17 (70.8) | 0.089 |
| Age, years | 62.6±13.1 | 64.1±12.0 | 61.4±14.7 | 58.6±14.6 | 0.176 |
| BMI, kg/m2 | 22.7±4.0 | 22.5±4.2 | 23.3±2.8 | 22.9±3.9 | 0.74 |
| Dialysis vintage, months | 31.0 (14.5, 48.0) | 32.0 (16.0, 48.0) | 34.5 (18.0, 49.5) | 19.5 (9.5, 39.5) | 0.279 |
| Vascular access | |||||
| AVF, n (%) | 104 (87.4) | 68 (88.3) | 17 (94.4) | 19 (79.2) | 0.303 |
| Catheter, n (%) | 15 (12.6) | 9 (11.7) | 1 (5.6) | 5 (20.8) | |
| Residual diuresis, mL/24 h | 100.0 (0, 300.0) | 100.0 (0, 300.0) | 150.0 (0, 200.0) | 200.0 (100.0, 700.0) | 0.084 |
| Interdialytic weight gain, kg | 2.2 (1.6, 3.0) | 2.0 (1.4, 2.7) | 2.6 (2.1, 3.3)** | 2.5 (1.6, 3.4) | 0.06 |
| History of smoking, n (%) | 37 (31.1) | 24 (31.2%) | 6 (33.3%) | 7(29.2%) | 0.959 |
| Hypertension, n (%) | 96 (80.7) | 60 (77.9) | 13 (72.2) | 23 (95.8) | 0.079 |
| Diabetes, n (%) | 61 (51.3) | 39 (50.6) | 5 (27.8) | 17 (70.8)* versus B2 | 0.022 |
| CVD, n (%) | 56 (47.1) | 25 (32.5) | 11 (61.1) | 20 (83.3)* | <0.001 |
| Hyperlipidaemia, n (%) | 51 (42.9) | 34 (44.2) | 8 (44.4) | 9 (37.5) | 0.818 |
| LVDD, n (%) | 57 (47.9) | 20 (26.0) | 14 (77.8) | 23 (95.8) | <0.001 |
| Haemoglobin, g/L | 115.6±16.9 | 118.6±14.2 | 112.7±13.8 | 107.9±23.7 | 0.059 |
| Albumin, g/L | 40.0±3.8 | 41.0±3.1 | 39.0±4.7 | 37.5±4.3** | <0.001 |
| CRP, mg/L | 3.7 (0.6, 11.7) | 1.7 (0.5, 5.6) | 6.0 (3.7, 13.6)** | 13.1 (6.4, 18.5)** | <0.001 |
| Calcium, mmol/L | 2.2±0.2 | 2.3±0.2 | 2.1±0.2 | 2.1±0.3* | 0.004 |
| Phosphate, mmol/L | 1.9±0.7 | 1.9±0.7 | 1.7±0.5 | 2.1±0.8 | 0.216 |
| iPTH, pg/mL | 185.0 (107.0, 295.0) | 195.0 (108.0, 301.0) | 183.5 (99.2, 282.0) | 158.0 (109.5, 228.6) | 0.614 |
| BNP, pg/mL | 230.7 (99.5, 510.9) | 164.3 (91.7, 368.8) | 375.9(99.0, 772.9) | 601.0 (353.6, 1,169.5)** | <0.001 |
| Echocardiographic parameters | |||||
| LVEF (%) | 62.0 (57.0, 65.0) | 63.0 (59.5, 65.0) | 63.5 (50.0, 67.0) | 57.0 (47.0, 60.0)** | 0.001 |
| LAVI, mL/m2 | 21.2 (17.6, 24.9) | 19.3 (16.9, 22.9) | 21.3 (19.6, 23.5) | 25.1 (24.2, 30.1)** | <0.001 |
| LVEDVI, mL/m2 | 310.9±107.7 | 283.1±88.2 | 348.8±119.3 | 371.5±125.6** | 0.004 |
| LVESVI, mL/m2 | 97.7±58.9 | 79.8±35.1 | 113.3±60.1 | 143.3±87.3** | 0.001 |
| LVMI, g/m2 | 111.5 (98.4, 135.0) | 107.6 (97.0, 122.3) | 124.4 (98.0, 138.2) | 129.5 (109.4, 157.6)** | 0.003 |
| e′, cm/s | 7.4 (6.0, 10.0) | 7.1 (6.0, 8.7) | 6.0 (5.0, 8.0) | 9.5 (7.8, 10.0)* | 0.012 |
| E/A | 0.8 (0.7, 1.1) | 0.7 (0.6, 0.9) | 0.8 (0.6, 1.0) | 1.2 (0.9, 1.6)** | <0.001 |
| E/e′ | 9.4 (8.0, 12.0) | 8.8 (7.1, 11.6) | 10.4 (10.0, 12.8)** | 10.0 (9.0, 14.0)** | 0.003 |
| Pulmonary pressure, mm Hg | 31.0 (25.0, 42.0) | 30.5 (25.0, 40.0) | 25.0 (22.0, 32.0) | 41.0 (31.5, 50.5)* | 0.001 |
Data are expressed as mean ± SD, median and interquartile range or as per cent frequency, as appropriate. Patients are divided into three groups on the basis of LUS B-line scores. p tests the differences among the groups. Significant differences between groups are indicated in bold. *p < 0.05, **p < 0.01 versus the B1.
BMI, body mass index; AVF, arterio-venous fistula; CVD, cardiovascular disease; LVDD, left ventricular diastolic dysfunction; CRP, C-reactive protein; iPTH, intact parathyroid hormone; BNP, B-type natriuretic peptide; LVEF, left ventricular ejection fraction; LAVI, left atrial volume index; LVEDVI, left ventricular end-diastolic volume index; LVESVI, left ventricular end-systolic volume index; LVMI, left ventricular mass index; e′, early diastolic mitral annular velocity; E/A, early to atrial filling velocity; E/e′, ratio of the early transmitral flow velocity to early diastolic mitral annular velocity.
Correlation Analysis between B-Lines and Clinical or Echocardiographic Parameters
Univariate correlation analysis between B-lines and clinical or echocardiographic parameters is summarized in Table 2. Table 3 presents the stepwise multivariate linear regression analysis for the predictors of echocardiographic parameters. After adjustment for age, sex, BMI, dialysis vintage, B-line, BNP, residual diuresis, vascular access, and comorbidities, including diabetes, CVD, hyperlipidaemia, and hypertension, the only variable expected to influence E/e′ was B-line (p = 0.002). The factors expected to influence the E/A ratio were B-line (p = 0.008) and residual diuresis (p = 0.036). The variables expected to influence LVEF were BNP (p < 0.001) and dialysis vintage (p = 0.044). Overall, the association of B-lines with echocardiographic parameters (LAVI, LVEDVI, E/A, and E/e′) of LV diastolic function was statistically significant (Table 3), and the correlation of B-lines with LVEF disappeared after adjustment for confounding factors.
Table 2.
Correlation analysis for the association of LUS B-lines with clinical and echocardiographic data
| Variables | B-lines | |
|---|---|---|
| r | p value | |
| Age (years) | −0.156 | 0.091 |
| BMI (kg/m2) | 0.043 | 0.643 |
| Dialysis vintage (months) | −0.141 | 0.135 |
| Residual diuresis (mL/24 h) | 0.156 | 0.091 |
| Interdialytic weight gain (kg) | 0.125 | 0.176 |
| Haemoglobin (g/L) | −0.224 | 0.014 |
| CRP (mg/L) | 0.428 | <0.001 |
| Albumin (g/L) | −0.392 | <0.001 |
| Calcium (mmol/L) | −0.252 | 0.006 |
| Phosphate (mmol/L) | 0.063 | 0.497 |
| iPTH (pg/mL) | −0.032 | 0.733 |
| BNP (pg/mL) | 0.316 | 0.001 |
| LVEF (%) | −0.284 | 0.002 |
| LAVI (mL/m2) | 0.339 | <0.001 |
| LVEDVI (mL/m2) | 0.220 | 0.016 |
| LVESVI (mL/m2) | 0.270 | 0.003 |
| LVMI (g/m2) | 0.271 | 0.003 |
| e′ (cm/s) | 0.086 | 0.361 |
| E/A | 0.456 | <0.001 |
| E/e′ | 0.296 | 0.001 |
| Pulmonary pressure (mm Hg) | 0.179 | 0.064 |
Data are expressed as correlation coefficient (r) and p value. Using Pearson or Spearman correlation analysis is based on the distribution and type of the data. Significant correlations are indicated in bold.
BMI, body mass index; CRP, C-reactive protein; iPTH, intact parathyroid hormone; BNP, B-type natriuretic peptide; LVEF, left ventricular ejection fraction; LAVI, left atrial volume index; LVEDVI, left ventricular end-diastolic volume index; LVESVI, left ventricular end-systolic volume index; LVMI, left ventricular mass index; e′, early diastolic mitral annular velocity; E/A, early to atrial filling velocity; E/e′, ratio of the early transmitral flow velocity to early diastolic mitral annular velocity.
Table 3.
Stepwise multivariate linear regression analysis for the predictors of echocardiographic parameters (E/e′, E/A, LAVI, LVEDVI, LVESVI, LVMI, LVEF, and pulmonary pressure)
| Variables | β | Standardized β | 95% CI | p value |
|---|---|---|---|---|
| E/e′ | ||||
| B-lines | 1.956 | 0.351 | 0.747–3.165 | 0.002 |
| E/A | ||||
| B-lines | 0.369 | 0.300 | 0.100–0.639 | 0.008 |
| Residual diuresis | 0.000 | 0.000 | 0.000–0.001 | 0.036 |
| LAVI (mL/m2) | ||||
| B-lines | 2.219 | 0.253 | 0.433–4.004 | 0.016 |
| BNP (pg/mL) | 2.472 | 0.266 | 0.439–4.506 | 0.018 |
| CVD versus no CVD | 2.260 | 0.237 | 0.292–4.229 | 0.025 |
| LVEDVI (mL/m2) | ||||
| B-lines | 55.049 | 0.281 | 12.626–97.471 | 0.012 |
| LVESVI (mL/m2) | ||||
| BNP (pg/mL) | 33.011 | 0.328 | 11.591–54.431 | 0.003 |
| LVMI (g/m2) | ||||
| BNP (pg/mL) | 19.206 | 0.311 | 5.966–32.446 | 0.005 |
| LVEF (%) | ||||
| BNP (pg/mL) | −6.725 | −0.431 | −9.880 to −3.571 | <0.001 |
| Dialysis vintage (months) | 0.038 | 0.209 | 0.001–0.074 | 0.044 |
| Pulmonary pressure (mm Hg) | ||||
| Age | 0.278 | 0.354 | 0.105–0.452 | 0.002 |
All the stepwise multivariate regression analysis models comprised the variables: age, sex, BMI, dialysis vintage, B-lines, BNP, residual diuresis, vascular access, and the comorbidities including diabetes, CVD, hyperlipidaemia, and hypertension. Category variables (sex, vascular access, and the comorbidities) were transformed into dummy variables through assignment codes before regression analysis. B-lines and BNP values were natural logarithm transformed.
β, regression coefficient; CI, confidence interval; BMI, body mass index; AVF, arterio-venous fistula; CVD, cardiovascular disease; BNP, B-type natriuretic peptide; LVEF, left ventricular ejection fraction; LAVI, left atrial volume index; LVEDVI, left ventricular end-diastolic volume index; LVESVI, left ventricular end-systolic volume index; LVMI, left ventricular mass index; e′, early diastolic mitral annular velocity; E/A, early to atrial filling velocity; E/e′, ratio of the early transmitral flow velocity to early diastolic mitral annular velocity.
Differences in B-Lines according to LAVI, LVEDVI, E/A, E/e′, and LVDD Status
Levels of B-lines were compared according to tertiles of LAVI, LVEDVI, E/A, and E/e′ and LVDD status using a multivariate analysis (Fig. 2). The highest tertile of the LAVI and E/A groups (LAVI3 and E/A3) had more B-lines than the first and second tertiles of the LAVI and E/A groups (LAVI1 and LAVI2, E/A1 and E/A2; Fig. 2a, c, respectively). Compared with the second tertile of the LVEDVI group (LVEDVI2) and the first tertile of the E/e′ group (E/e′1), the level of B-lines increased in the highest tertile of the LVEDVI group (LVEDVI3) and in the E/e′ group (E/e′3) (Fig. 2b, d). The LVDD group had more B-lines than the non-LVDD group (Fig. 2e).
Fig. 2.
Differences in LUS B-lines according to tertiles of LAVI, LVEDVI, E/A, and E/e′, and LVDD status. Levels of LUS B-lines were compared according to tertiles of LAVI, LVEDVI, E/A, and E/e′, and status of LVDD using a multivariate analysis. The LAVI3 group had more B-lines than LAVI1 and LAVI2 groups (a). Compared with the LVEDVI2 group, the level of B-lines increased in the LVEDVI3 group (b). The level of B-lines was higher in the E/A3 group than that in E/A1 and E/A2 groups (c). Furthermore, participants in the E/e′3 group had significantly more B-lines than those in the E/e′1 group (d). The LVDD group had more B-lines than the non-LVDD group (e). a B-lines in LAVI1 (LAVI <18.93 mL/m2), LAVI2 (18.93 mL/m2 ≤ LAVI ≤23.55 mL/m2), and LAVI3 (LAVI >23.55 mL/m2). b B-lines in LVEDVI1 (LVEDVI <251.96 mL/m2), LVEDVI2 (251.96 mL/m2 ≤ LVEDVI ≤340.67 mL/m2), and LVEDVI3 (LVEDVI >340.67 mL/m2). c B-lines in E/A1 (E/A <0.69), E/A2 (0.69≤ LVEDVI ≤0.88), and E/A3 (E/A >0.88). d B-lines in E/e′1 (E/e′ <8.36), E/e′2 (8.36≤ E/e′ ≤11.20), and E/e′3 (E/e′ >11.20). e B-lines in non-LVDD and LVDD. *p < 0.05, **p < 0.01 versus LAVI1, LVEDVI1, E/A1, E/e′1, non-LVDD, respectively. #p < 0.05, ##p < 0.01 versus LAVI2, LVEDVI2, E/A2, E/e′2, respectively. LAVI, left atrial volume index; LVEDVI, left ventricular end-diastolic volume index; E/A, early to atrial filling velocity; E/e′, ratio of the early transmitral flow velocity to early diastolic mitral annular velocity; LVDD, left ventricular diastolic dysfunction.
Logistic Regression Analyses for Predictors of LVDD Based on B-Lines and Clinical Parameters
Table 4 shows the results of the final model using the forward stepwise method of logistic regression consisting of B-line category, sex, CVD, and diabetes. Compared with the mild B-line group (B1: <14), the moderate and severe B-line groups (B2: 14–30; B3: >30, respectively) were approximately 24 and 95 times more likely to have LVDD, respectively (OR 24.344 [95% CI 4.854–122.084], p < 0.001, and OR 94.552 [9.617–929.022], p < 0.001). Compared with the male group, the female group (OR 4.202 [1.249–14.145], p = 0.02) had an increased LVDD risk. The presence of CVD (OR 3.660 [1.229–10.901], p = 0.02) and diabetes (OR 4.602 [1.380–15.346], p = 0.013) was associated with an increased risk of LVDD.
Table 4.
Multivariate analysis using logistic regression: predictive factors of LVDD
| Variables | β | SE | Wald | p value | OR (95% CI) |
|---|---|---|---|---|---|
| B-line categories | |||||
| B1 (mild: <14) | Reference | Reference | Reference | Reference | Reference |
| B2 (moderate: 14–30) | 3.192 | 0.823 | 15.057 | <0.001 | 24.344 (4.854–122.084) |
| B3 (severe: >30) | 4.549 | 1.166 | 15.220 | <0.001 | 94.522 (9.617–929.022) |
| Sex | |||||
| Male | Reference | Reference | Reference | Reference | Reference |
| Female | 1.436 | 0.619 | 5.375 | 0.02 | 4.202 (1.249–14.145) |
| CVD | |||||
| No CVD | Reference | Reference | Reference | Reference | Reference |
| CVD | 1.297 | 0.557 | 5.430 | 0.02 | 3.660 (1.229–10.901) |
| Diabetes | |||||
| No diabetes | Reference | Reference | Reference | Reference | Reference |
| Diabetes | 1.527 | 0.614 | 6.173 | 0.013 | 4.602 (1.380–15.346) |
The logistic regression analysis models comprised the variables: age, sex, BMI, dialysis vintage, B-lines, residual diuresis, vascular access, and the comorbidities including diabetes, CVD, hyperlipidaemia, and hypertension. All the variables were subjected to forward stepwise (likelihood ratio) logistic regression using Hosmer-Lemeshow test.
β, regression coefficient; SE, standard error; OR, odds ratio; CI, confidence interval; CVD, cardiovascular disease; LVDD, left ventricular diastolic dysfunction.
ROC Curve for Assessing the Value of B-Lines as a Predictor of LVDD
The area under the curve of the ROC curve for B-lines predicting LVDD among clinically euvolemic haemodialysis patients was 0.845 (95% CI 0.771–0.919, p < 0.0001; Fig. 3). The cut-off of B-lines for predicting LVDD was 14.5, with a sensitivity and specificity of 64.91% and 93.55%, respectively.
Fig. 3.
ROC curve for assessing the value of LUS B-lines predicting LVDD among clinically euvolemic haemodialysis patients, The AUC of the ROC curve for LUS B-lines predicting LVDD among clinically euvolemic haemodialysis patients was 0.845 (SE 0.038, 95% CI 0.771–0.919, p < 0.0001). The cut-offs of B-lines for predicting LVDD were 14.5 with a sensitivity and specificity of 64.91% and 93.55%, respectively. ROC, receiver operating characteristic; AUC, area under the curve; LVDD, left ventricular diastolic dysfunction; CI, confidence interval; SE, standard error.
Discussion
To the best of our knowledge, this is the first study to have examined the association of B-lines with LVDD by the 2016 ASE/EACVI recommendations [15] in clinically euvolemic haemodialysis patients. Our study demonstrated that B-line number was statistically related to echocardiographic parameters of diastolic function but was not correlated with LVEF. Additionally, compared with the mild B-line group, the moderate and severe B-line groups were correlated with an increased risk of LVDD. Furthermore, the cut-off of B-lines for predicting LVDD was 14.5. Taken together, these results suggest the usefulness of LUS for evaluating LVDD in haemodialysis patients.
In this study, the prevalence of LVDD based on the new 2016 ASE/EACVI recommendations was 47.9% in clinically euvolemic haemodialysis patients, which was significantly higher than the prevalence of 1.4% in the general population reported by a recent study [19]. Our observation conforms to those of previous studies that showed that the prevalence of LVDD varies from 25% to 87% in haemodialysis patients [3, 4]. This high prevalence may be related to the presence of traditional risk factors and to the synergistic effects of these factors associated with ESRD and dialysis treatment [20]. Accordingly, our study showed that female sex and the presence of CVD and diabetes were all associated with an increased risk of LVDD.
LUS is a well-validated, low-cost, and simple technique that can be easily applied by nephrologists [21]. LUS through B-line evaluation can be used to assess ELW determined by the filling pressure of the LV [9, 22]. An earlier study showed that 57% of asymptomatic haemodialysis patients had moderate to severe ELW [11]. In our study, the percentage was 35.3%. A possible explanation is that only the clinically euvolemic haemodialysis population was included in the present study, thereby reducing the effect of volume load on ELW. Moreover, five studies have assessed the relationship between B-lines and different echocardiographic parameters in haemodialysis patients [11, 13, 23–25]. Mallamaci et al. [11] demonstrated for the first time in a haemodialysis population that B-lines were correlated to anatomical and functional echocardiographic parameters. Multiple regression analysis revealed that only LVEF maintained an independent link with B-lines. In line with this, another study [25] showed that B-lines were negatively correlated with LVEF and were positively related to mitral gradient. In contrast, Siriopol et al. [13] and Donadio et al. [23] did not reveal significant relationships between predialysis B-lines and any echocardiographic parameters. Only postdialysis B-lines were found to be related to left atrial diameter and LVEF [13] or to E/e′, LVEF, and pulmonary pressure [23]. Interestingly, the two studies did not confirm the association between the change in B-line number and any echocardiographic characteristic [13, 24], while another study [26] indicated that B-line-guided dry weight reduction is associated with reverse LVEDVI and LAVI, not with LVEF, suggesting that the reduction in B-line number may improve LV diastolic filling properties. In our study, B-lines were statistically related to LAVI, LVEDVI, E/A, and E/e′ but were not correlated with LVEF after adjustments for confounding factors. Such discrepancies may be attributed to differences in the studied population and the degree to which the patients were cardiovascularly compromised.
To date, there is no report about the relationship between B-lines and LVDD assessed comprehensively by recommendations of echocardiographic algorithms instead of a single indicator such as E/A or E/e′ in clinically euvolemic haemodialysis patients. Notably, we revealed that the moderate and severe B-line groups were correlated with an increased risk of LVDD. Moreover, our findings suggested a B-line ≥14.5 value as a reliable cut-off for identifying LVDD in haemodialysis patients. These data support the use of LUC for the purpose of early detection of LVDD. Additionally, LUS is more convenient, rapid, and low cost compared with a full echocardiography examination.
The present study has several limitations. Firstly, the sample size was relatively small, and the cross-sectional nature of the present study makes it difficult to determine any causal relationship. Secondly, it was conducted at a single centre, suggesting the possibility of selection bias. Therefore, the results should be interpreted with caution, and future large studies need to be conducted to demonstrate the diagnostic value of B-lines for LVDD.
Conclusions
Our study showed that B-lines were closely associated with LV diastolic function in clinically euvolemic haemodialysis patients. Moreover, our findings suggested a B-line ≥14.5 value as a reliable cut-off in identifying patients with LVDD. LUS B-lines may be used as a novel indicator for evaluating LVDD.
Acknowledgments
The authors thank all the staff of the haemodialysis department for the practical support and the Ethics Committee of the Tongren Hospital Affiliated to Shanghai Jiao Tong University.
Statement of Ethics
The study protocol was reviewed and approved by the Ethics Committee of the Tongren Hospital Affiliated to Shanghai Jiao Tong University, approval number 2021-008-01. Written informed consent was obtained from participants to participate in the study.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This work was supported by Science and Technology Committee Foundation of Shanghai Changning District (Grant No. CNKW2020Y09), the National Natural Science Foundation of China (Grant No. 82100766, 82170745, and 82000687), and the Research Fund of Shanghai Tongren Hospital, Shanghai Jiao Tong University School of Medicine (Grant No. TRGG202106 and TRKYRC-xx202206).
Author Contributions
Xiaoxia Wang and Fengqin Li contributed to the research idea and study design. Fengqin Li was responsible for original draft preparation and performed the statistical analysis. Fengqin Li, Miao Ding, and Yanzhe Wang carried out the research. Fengqin Li, Miao Ding, Qijie Chen, Yanzhe Wang, Yue Wu, Chuchu Zeng, Nan Zhang, and Dingyu Zhu contributed to data recording and acquisition. All authors contributed to clinical care of patients, commented on previous versions of the manuscript, and read and approved the final manuscript.
Funding Statement
This work was supported by Science and Technology Committee Foundation of Shanghai Changning District (Grant No. CNKW2020Y09), the National Natural Science Foundation of China (Grant No. 82100766, 82170745, and 82000687), and the Research Fund of Shanghai Tongren Hospital, Shanghai Jiao Tong University School of Medicine (Grant No. TRGG202106 and TRKYRC-xx202206).
Data Availability Statement
The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants, but are available from the corresponding author X.W. upon reasonable request.
References
- 1. de Jager DJ, Grootendorst DC, Jager KJ, van Dijk PC, Tomas LM, Ansell D, et al. Cardiovascular and noncardiovascular mortality among patients starting dialysis. JAMA. 2009;302(16):1782–9. [DOI] [PubMed] [Google Scholar]
- 2. Saran R, Robinson B, Abbott KC, Agodoa L, Bragg-Gresham J, Balkrishnan R, et al. US renal data system 2018 annual data report: epidemiology of kidney disease in the United States. Am J Kidney Dis. 2019;73(3 Suppl 1):A7–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Chiu DY, Green D, Abidin N, Sinha S, Kalra PA. Echocardiography in hemodialysis patients: uses and challenges. Am J Kidney Dis. 2014;64(5):804–16. [DOI] [PubMed] [Google Scholar]
- 4. Assa S, Hummel YM, Voors AA, Kuipers J, Groen H, de Jong PE, et al. Changes in left ventricular diastolic function during hemodialysis sessions. Am J Kidney Dis. 2013;62(3):549–56. [DOI] [PubMed] [Google Scholar]
- 5. Han JH, Han JS, Kim EJ, Doh FM, Koo HM, Kim CH, et al. Diastolic dysfunction is an independent predictor of cardiovascular events in incident dialysis patients with preserved systolic function. PLoS One. 2015;10(3):e0118694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Barberato SH, Bucharles SG, Sousa AM, Costantini CO, Costantini CR, Pecoits-Filho R. Prevalência e impacto prognóstico da disfunção diastólica na doença renal crônica em hemodiálise. Arq Bras Cardiol. 2010;94(4):457–62. [DOI] [PubMed] [Google Scholar]
- 7. Ogawa T, Nitta K. Clinical impact of left ventricular diastolic dysfunction in chronic kidney disease. Contrib Nephrol. 2018;195:81–91. [DOI] [PubMed] [Google Scholar]
- 8. Hubert A, Girerd N, Le Breton H, Galli E, Latar I, Fournet M, et al. Diagnostic accuracy of lung ultrasound for identification of elevated left ventricular filling pressure. Int J Cardiol. 2019;281:62–8. [DOI] [PubMed] [Google Scholar]
- 9. Covic A, Siriopol D, Voroneanu L. Use of lung ultrasound for the assessment of volume status in CKD. Am J Kidney Dis. 2018;71(3):412–22. [DOI] [PubMed] [Google Scholar]
- 10. Volpicelli G, Elbarbary M, Blaivas M, Lichtenstein DA, Mathis G, Kirkpatrick AW, et al. International evidence-based recommendations for point-of-care lung ultrasound. Intensive Care Med. 2012;38(4):577–91. [DOI] [PubMed] [Google Scholar]
- 11. Mallamaci F, Benedetto FA, Tripepi R, Rastelli S, Castellino P, Tripepi G, et al. Detection of pulmonary congestion by chest ultrasound in dialysis patients. JACC Cardiovasc Imaging. 2010;3(6):586–94. [DOI] [PubMed] [Google Scholar]
- 12. Zoccali C, Torino C, Tripepi R, Tripepi G, D’Arrigo G, Postorino M, et al. Pulmonary congestion predicts cardiac events and mortality in ESRD. J Am Soc Nephrol. 2013;24(4):639–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Siriopol D, Hogas S, Voroneanu L, Onofriescu M, Apetrii M, Oleniuc M, et al. Predicting mortality in haemodialysis patients: a comparison between lung ultrasonography, bioimpedance data and echocardiography parameters. Nephrol Dial Transpl. 2013;28(11):2851–9. [DOI] [PubMed] [Google Scholar]
- 14. Frassi F, Gargani L, Gligorova S, Ciampi Q, Mottola G, Picano E. Clinical and echocardiographic determinants of ultrasound lung comets. Eur J Echocardiogr. 2007;8(6):474–9. [DOI] [PubMed] [Google Scholar]
- 15. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF 3rd, Dokainish H, Edvardsen T, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American society of echocardiography and the European association of cardiovascular imaging. J Am Soc Echocardiogr. 2016;29(4):277–314. [DOI] [PubMed] [Google Scholar]
- 16. Loutradis C, Sarafidis PA, Ekart R, Papadopoulos C, Sachpekidis V, Alexandrou ME, et al. The effect of dry-weight reduction guided by lung ultrasound on ambulatory blood pressure in hemodialysis patients: a randomized controlled trial. Kidney Int. 2019;95(6):1505–13. [DOI] [PubMed] [Google Scholar]
- 17. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American society of echocardiography and the European association of, cardiovascular imaging. Eur Heart J Cardiovasc Imaging. 2016;17(4):412. [DOI] [PubMed] [Google Scholar]
- 18. Jambrik Z, Monti S, Coppola V, Agricola E, Mottola G, Miniati M, et al. Usefulness of ultrasound lung comets as a nonradiologic sign of extravascular lung water. Am J Cardiol. 2004;93(10):1265–70. [DOI] [PubMed] [Google Scholar]
- 19. Almeida JG, Fontes-Carvalho R, Sampaio F, Ribeiro J, Bettencourt P, Flachskampf FA, et al. Impact of the 2016 ASE/EACVI recommendations on the prevalence of diastolic dysfunction in the general population. Eur Heart J Cardiovasc Imaging. 2018;19(4):380–6. [DOI] [PubMed] [Google Scholar]
- 20. Ogawa T, Koeda M, Nitta K. Left ventricular diastolic dysfunction in end-stage kidney disease: pathogenesis, diagnosis, and treatment. Ther Apher Dial. 2015;19(5):427–35. [DOI] [PubMed] [Google Scholar]
- 21. Picano E, Gargani L, Gheorghiade M. Why, when, and how to assess pulmonary congestion in heart failure: pathophysiological, clinical, and methodological implications. Heart Fail Rev. 2010;15(1):63–72. [DOI] [PubMed] [Google Scholar]
- 22. Agricola E, Bove T, Oppizzi M, Marino G, Zangrillo A, Margonato A, et al. “Ultrasound comet-tail images”: a marker of pulmonary edema: a comparative study with wedge pressure and extravascular lung water. Chest. 2005;127(5):1690–5. [DOI] [PubMed] [Google Scholar]
- 23. Donadio C, Bozzoli L, Colombini E, Pisanu G, Ricchiuti G, Picano E, et al. Effective and timely evaluation of pulmonary congestion: qualitative comparison between lung ultrasound and thoracic bioelectrical impedance in maintenance hemodialysis patients. Medicine. 2015;94(6):e473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Weitzel WF, Hamilton J, Wang X, Bull JL, Vollmer A, Bowman A, et al. Quantitative lung ultrasound comet measurement: method and initial clinical results. Blood Purif. 2015;39(1–3):37–44. [DOI] [PubMed] [Google Scholar]
- 25. Pardała A, Lupa M, Chudek J, Kolonko A. Lung ultrasound B-lines occurrence in relation to left ventricular function and hydration status in hemodialysis patients. Medicina. 2019;55(2):45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Loutradis C, Papadopoulos CE, Sachpekidis V, Ekart R, Krunic B, Papadopoulou D, et al. Lung ultrasound-guided dry-weight reduction and echocardiographic changes in clinically euvolemic hypertensive hemodialysis patients: 12-month results of a randomized controlled trial. Hellenic J Cardiol. 2022;64:1–6. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants, but are available from the corresponding author X.W. upon reasonable request.
