Rhythm outcomes after aortic valve surgery: Treatment and evolution of new‐onset atrial fibrillation

1 INTRODUCTION

New-onset atrial fibrillation (AF) after cardiac surgery is a common complication that is associated with major adverse events.1, 2 It has been well reported that new-onset AF after aortic valve (AV) procedures elevates in-hospital mortality, while its sole effect on long-term endpoints remain controversial.3-5 It should be noted that most of the studies attempted to relate new-onset AF to late events directly, without detailed heart rhythm follow-ups to support the causal relationship.

Regarding the optimal management of postoperative new-onset AF, current guidelines and recent trials recommend both rate control and rhythm control therapies.1, 6 At our institution, rhythm control is the preferred choice, and conversion is promptly performed for new-onset AF. However, there is a lack of clinical and rhythm data to validate the mid-term effectiveness of such strategy. Furthermore, in addition to the well-known predictors of postoperative new-onset AF (e.g., age and left atrial size), risk factors that may predispose patients to long-term AF after AV surgery are also clinically relevant.

Hence, in this study, we aimed to (1) follow up heart rhythm after AV surgery; (2) compare clinical and rhythm outcomes between those with rhythm controlled new-onset AF and those who remained in sinus rhythm throughout postoperative hospitalization; and (3) identify predictors of paroxysmal or persistent AF at 1 year.

2 PATIENTS AND METHODS 2.1 Study population

This study was approved by the Ethics Committee of Patients of Zhongshan Hospital. The requirement for informed consent was waived because of the retrospective nature of the study (approval number: B2021-534R). Between January 2017 and November 2018, data of 3667 consecutive patients undergoing AV surgery at our Department of Cardiac Surgery (Zhongshan Hospital Fudan University, Shanghai, China) were reviewed. We excluded patients with histories of AF, atrial flutter or atrial tachycardia, hyperthyroidism, and those who underwent cardiac reoperations. Concomitant procedures were limited to root reconstruction and ascending aortic repair. Patients undergoing transapical transcatheter AV implantation were also excluded. After screening, 978 patients were selected as the study cohort (Figure 1).

image

Flowchart presenting the inclusion criteria and rhythm outcomes. AF, atrial fibrillation

2.2 Operative procedures

All patients underwent AV replacement or repair, with concomitant procedures limited to root reconstruction and ascending aortic repair. Patients with concomitant coronary artery, other valve or aortic arch procedures were excluded. The details of surgical procedures were reported in previous studies.7, 8 Invasive approaches included sternotomy (n = 871, 89.1%), supra-sternotomy (n = 79, 8.1%) and minimally invasive right thoracotomy (n = 28, 2.9%). The left ventricle was routinely vented via right superior pulmonary vein. Bioprostheses were implanted in 360 (36.8%) patients. No patient received prophylactic pulmonary vein isolation or ligation of left atrial appendage.

2.3 New-onset AF: Rhythm control therapy

At our institution, we did not prophylactically use Class III antiarrhythmic drugs. Instead, preoperative beta-blockers were routinely administered in all patients planned for AV surgery, unless contraindicated. The heart rhythm of each patient was monitored continuously from postoperative day 0 until discharge using telemetry. New-onset AF was defined as postoperative AF that lasted at least for 30 s or recurred during hospitalization.9 Whenever a period of AF was noted by the nursing staff, the physician on-call would be informed and respond accordingly. Data of the rhythm control treatment, including intravenous and/or oral use of Class III antiarrhythmic drugs with and without beta-blockers and electrical cardioversion, were collected by interrogating records of daily rounds and order lists.

Warfarin therapy was initiated on postoperative day 0 in all patients, and doses were titrated to achieve an international normalized ratio of 2.0–3.0. Patients with new-onset AF that was not successfully converted were discharged on continuous oral amiodarone or sotalol for 3 months. Antiarrhythmic drugs were discontinued in patients with completed 3-month regimen and in those who developed drug-related side effects, including dizziness, blurred vision, cough, dyspnea, symptomatic bradycardia, electrocardiogram-confirmed long-QT syndrome or ventricular tachycardia.

Heart rhythm follow-ups included electrocardiogram, 24-h Holter monitoring and pacemaker interrogation, which were performed at the discretion of the referring cardiologist and the operating surgeon. Specifically, electrocardiogram and pacemaker interrogation were repeated at 1, 3, 6, and 12 months after surgery, and yearly thereafter. Patients who complained of paroxysmal/continuous palpitations received 24-h Holter monitoring. Additional Holter monitoring was performed if previous reports did not reveal AF while the symptoms persisted.

For patients who underwent bioprosthetic AV replacement or valve repair, anticoagulation was discontinued if sinus rhythm was stable at 3 months. If AF was still detectable at 3 months, electrical cardioversion or transcatheter ablation was recommended. Rate control and anticoagulation therapies were used in patients with persistent AF who had undergone at least 1 electrical cardioversion or transcatheter ablation, and in those who refused those procedures.

2.4 Follow-up

The primary outcomes of this study were all-cause mortality and stroke. The secondary outcomes were heart rhythm statuses at 3 months and at 1 year after surgery. Data of patient status and therapy after discharge were prospectively collected via telephone calls and the outpatient clinic database. Collection of follow-up data was performed between November 1, 2019 and January 15, 2020. Clinical follow-up was 93.9% (918/978) complete at a median of 19 months (interquartile range, 14–25 months). The baseline and perioperative data were compared between patients with and without follow-up data (Table S1).

Heart rhythm data were available in 93.8% (n = 907) survivors at 3 months, and in 92.1% (n = 836) survivors at 1 year, including a total of 2588 electrocardiograms, 65 Holter monitoring reports and 30 pacemaker interrogations during follow-up (median time [interquartile range], 15 [12–24] months). All patients with rhythm follow-ups had at least two electrocardiograms. In this study, data of survival and rhythm outcomes were collected by two independent investigators (B.X. and S.Y.) who were blinded to the baseline and perioperative data.

2.5 Statistical analysis

Continuous variables were presented as mean ± SD and compared using the Student t-test or Mann–Whitney U-test according to the Shapiro–Wilk normality test. Categorical variables were described as numbers and percentages and analyzed using the chi-square test or Fisher's exact test, as appropriate. The baseline and perioperative data were complete in all patients.

Multivariate logistic regression models were used to identify predictors of new-onset AF with the forward stepwise selection method (Pentry = .10, Pstay = .05). The baseline and operative covariates included in the models were age, gender, comorbidities, New York Heart Association functional class III–IV, baseline left atrial dimension, left ventricular ejection fraction, AV pathologies, operating procedures (surgery vs. intervention, valve repair vs. replacement and concomitant procedures), type and size of prosthesis, in-hospital morbidities, transfusion, hypokalemia, and perioperative inotropic agents. Covariates with P < .10 in univariate models were included in the multivariate model. The Hosmer–Lemeshow goodness-of-fit test and calculation of c-statistic were performed for the final model (model 1).

Kaplan–Meier curves were used to describe freedom from death and stroke. To compare survival between patients with and without new-onset AF, the inverse probability weighting method was used to adjust multiple covariates, including age, gender, coronary artery disease, diabetes, hypertension, chronic lung disease, and cerebrovascular disease, New York Heart Association functional class III–IV, left atrial dimension, left ventricular ejection fraction, surgical approaches and procedures. Competing risks of mortality and stroke were analyzed using the Fine–Gray method.

Multivariate logistic regression models were also established to investigate predictors of AF at 1 year. Apart from the covariates listed above, non-antiarrhythmic medications that were administered after 3 months, including loop diuretics, spironolactone, beta-blockers, angiotensin-converting enzyme inhibitor/angiotensin receptor blockers, digoxin and statins, were also included in the models (model 2).

A two-tailed P value <.05 was considered statistically significant. All statistical analyses were performed using R v3.3.3 (Package “IPWsurvival”, R Development Core Team, Vienna, Austria) and STATA 15 (StataCorp LP, TX, USA).

3 RESULTS 3.1 Patient demographics

The baseline, operative and postoperative characteristics were listed in Tables 1 and 2. The overall mean age was 59 years (range, 18–90 years) with 68.5% male patients. The overall in-hospital rates of mortality and stroke were 0.6% (n = 6) and 1.4% (n = 14), respectively. Postoperative new-onset AF occurred in 256 (26.2%) patients, which was associated with higher risks of in-hospital mortality (1.6% vs. 0.3%, P = .044). Rate of in-hospital stroke was comparable (2.3% vs. 1.1%, P = .22).

TABLE 1. Baseline and perioperative characteristics of the entire cohort and comparisons between patients with and without new-onset AF Demographics All (N = 978) No AF (n = 722) AF (n = 256) P value Male sex 670 (68.5) 486 (67.3) 184 (71.9) .18 Age 59 (48–66) 56 (46–64) 64 (55–69) <.001 Body mass index (kg/m2) 23.7 (21.8–25.0) 23.4 (21.0–24.2) 24.3 (22.5–26.0) .11 Diabetes 69 (7.1) 41 (5.7) 28 (10.9) .005 Hypertension 415 (42.4) 283 (39.2) 132 (51.6) <.001 Coronary artery disease 75 (7.7) 47 (6.5) 28 (10.9) .022 Chronic lung disease 25 (2.6) 14 (1.9) 11 (4.3) .040 Cerebrovascular disease 45 (4.6) 33 (4.6) 12 (4.7) .94 Chronic kidney disease 23 (1.8) 14 (1.9) 4 (1.6) >.99a Peripheral artery disease 16 (1.6) 8 (1.1) 8 (3.1) .042a Connective tissue disorder 17 (1.7) 15 (2.1) 2 (0.8) .26a Autoimmune disease 23 (2.4) 18 (2.5) 5 (2.0) .62 New York Heart Association functional class III–IV 734 (75.1) 527 (73.0) 207 (80.9) .012 Heart rate (beats per minute) 69 (63–77) 69 (62–77) 69 (63–79) .39 Left atrial dimension (mm) 40 (37–44) 40 (36–44) 42.8 ± 5.7 <.001 Left ventricular ejection fraction (%) 62 (56–66) 62 (57–66) 60 (51–65) <.001 Ejection fraction <50% 121 (12.4) 69 (9.6) 52 (20.3) <.001 Bicuspid AV 428 (43.8) 329 (45.6) 99 (38.7) .056 Aortic stenosis 181 (18.5) 129 (17.9) 52 (20.3) .39 Aortic regurgitation 539 (55.1) 406 (56.2) 133 (52.0) .24 Aortic stenosis + regurgitation 258 (26.4) 187 (25.9) 71 (27.7) .57 Mitral regurgitation >mild 22 (2.2) 15 (2.1) 7 (2.7) .54 Tricuspid regurgitation >mild 15 (1.5) 9 (1.3) 6 (2.3) .24a Ascending aortic diameter > 40 mm 521 (53.3) 374 (51.8) 147 (57.4) .12 EuroSCORE II (%) 1.34 ± 0.26 1.28 ± 0.20 1.51 ± 0.33 <.001 Note: Continuous variables are presented as medians (interquartile ranges) or means ± SD, according to the normality test. Categorical variables are presented as numbers (percentages). Abbreviations: AF, atrial fibrillation; AV, aortic valve. a Fisher's exact test. TABLE 2. Operative and postoperative characteristics of the entire cohort and comparisons between patients with and without new-onset AF Demographics All (N = 978) No AF (n = 722) AF (n = 256) P value Cardiopulmonary bypass time (min) 91 (70–111) 89 (68–108) 101.5 ± 31.7 .070 Aortic cross-clamp time (min) 57 (45–77) 56 (45–75) 58 (47–82) .23 Approaches Sternal 871 (89.1) 629 (87.1) 242 (94.5) .001 Supra-sternal 79 (8.1) 67 (9.3) 12 (4.6) .021 Right-thoracic 28 (2.9) 26 (3.6) 2 (0.8) .020 Bioprosthesis 360 (36.8) 222 (30.8) 138 (53.9) <.001 Prosthetic size 23 (21–25) 23 (21–25) 23 (21–25) .28 AV repair 62 (6.3) 53 (7.3) 9 (3.5) .031 Effective orifice area index <0.85 cm2/m2 111 (11.3) 77 (10.7) 34 (13.3) .26 Ascending aortic/root replacement 347 (35.5) 250 (34.6) 97 (37.9) .35 In-hospital mortality 6 (0.6) 2 (0.3) 4 (1.5) .044a Length of stay in the intensive care unit Morbidities 51 (5.2) 34 (4.7) 17 (6.6) .23 Low cardiac output syndrome 11 (1.1) 8 (1.1) 3 (1.2) >.99a Stroke 14 (1.4) 8 (1.1) 6 (2.3) .22a Dialysis 6 (0.6) 1 (0.1) 5 (1.9) .006a Ventilator support >96 h 17 (1.7) 9 (1.3) 8 (3.1) .089a Reoperation for bleeding 2 (0.2) 2 (0.3) 0 (0) >.99a Pacemaker implantation 17 (1.7) 14 (1.9) 2 (0.8) .26a Hypokalemia 185 (18.9) 112 (15.5) 73 (28.5) <.001 Transfusion 233 (23.8) 155 (21.5) 78 (30.2) .004 Red blood cell (IU) 0 (0–0) 0 (0–0) 0 (0–2) .002 Plasma (ml) 0 (0–0) 0 (0–0) 0 (0–400) .003 Perioperative medications Epinephrine/norepinephrine 876 (89.6) 643 (89.1) 233 (91.0) .38 Phosphodiesterase inhibitor 922 (94.3) 682 (94.5) 240 (93.8) .67 Dopamine/dobutamine 330 (33.7) 208 (28.8) 122 (47.7) <.001 Levosimendan 49 (5.0) 31 (4.3) 18 (7.0) .085 Note: Continuous variables are presented as medians (interquartile ranges) or means ± SD, according to the normality test. Categorical variables are presented as numbers (percentages). Abbreviations: AF, atrial fibrillation; AV, aortic valve. a Fisher's exact test. 3.2 New-onset AF and predictors

Treatments of new-onset AF included Class III antiarrhythmic drugs (n = 256), beta-blockers (n = 17) and electric cardioversion (n = 4) to restore sinus rhythm. As a result, 235 (93.3%) patients were successfully converted to sinus rhythm before discharge.

The final multivariate logistic model (model 1) showed that age per year (odds ratio [OR], 1.05; 95% confidence intervals [CI], 1.04–1.07; P < .001); baseline left atrial dimension per mm (OR, 1.06; 95% CI, 1.03–1.09; P < .001), left ventricular ejection fraction per % (OR, 0.98; 95% CI, 0.97–1.00; P = .045), hypokalemia (OR, 1.82; 95% CI, 1.26–2.62; P = .001) and perioperative use of dopamine/dobutamine (OR, 1.81; 95% CI, 1.32–2.48; P < .001) were significantly associated with occurrence of new-onset AF (c-statistic = 0.740; Hosmer–Lemeshow test, P = .47).

3.3 Primary outcomes

For the entire cohort, 14 (1.4%) deaths occurred after discharge, including 3 due to cerebrovascular events, 5 due to cardiac causes and 6 due to other causes. Stroke occurred in 15 (1.5%) patients after discharge. The 1-year cumulative rates of mortality and stroke were 1.7 ± 0.4% and 2.7 ± 0.5%, respectively. Before adjustment, there was significant differences in mortality and stroke between patients with and without new-onset AF (1-year mortality, 3.9 ± 0.1% vs. 0.7 ± 0.03%; 1-year rate of stroke, 4.0 ± 0.1% vs. 2.2 ± 0.1%; both P < .001; Figure 2). After adjustment, difference in stroke rate was still significant (subdistribution hazard ratio, 3.46; 95% CI, 1.22–9.84; P = .020), while mortality was comparable between patients with and without new-onset AF (adjusted log-rank, 1.13; P = .30).

image

Kaplan–Meier curves presenting differences in survival (A) and freedom from stroke (B) between patients with and without new-onset AF. AF, atrial fibrillation; CI, confidence interval

3.4 Rhythm follow-up and predictors of AF

Details of rhythm follow-ups and medications during follow-up were shown in Figure 1 and in Table 3, respectively. At 3 months, among 907 survivors with rhythm follow-ups, 21 (2.3%) patients had paroxysmal or persistent AF, including 12 (5.1%) patients with postoperative new-onset AF and 9 (1.3%) patients without (P < .001). Anticoagulation was continued in all those patients. Electric cardioversion and transcatheter ablation were performed in 5 and 2 patients with new-onset AF, respectively. Class III antiarrhythmic drugs were administered in the 9 patients without postoperative new-onset AF. Altogether, conversion was successful in 7 (33.3%) patients who were free from AF recurrence thereafter. At 1 year, among 836 survivors with rhythm follow-ups, the rate of persistent AF was 3.5% (n = 29), including 16 (7.5%) patients with postoperative new-onset AF and 13 (2.1%) patients without (P < .001). Patients were treated with long-term anticoagulants after ineffective rhythm control therapy. Among patients undergoing bioprosthetic AV replacement or valve repair, prolonged anticoagulation was administered in 11 (2.7%) patients at 3 months and in 14 (3.8%) patients at 1 year. Adjusted models did not reveal significant difference in rates of mortality and stroke between patients receiving mechanical prostheses and those with bioprostheses or repaired AV (adjusted P = .67 and 0.20, respectively).

TABLE 3. Postoperative medications after AV surgery in patients with and without new-onset AF Drugs 3 months (n = 907) 1 year (n = 836) No AF (n = 673) AF (n = 234) P value No AF (n = 623) AF (n = 213) P value Warfarin/aspirin therapy 78 (11.6) 44 (18.8) .005 80 (12.8) 41 (19.2) .022 Loop diuretics 653 (97.0) 225 (96.2) .51 47 (7.5) 22 (10.3) .20 Spironolactone 552 (82.0) 189 (80.8) .67 80 (12.8) 34 (16.0) .25 Beta-blockers 513 (76.2) 158 (67.5) .009 391 (62.8) 142 (66.7) .31 Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker 54 (8.0) 19 (8.1) .96 67 (10.8) 18 (8.5) .37 Digoxin 5 (0.7) 3 (1.3) .43a 9 (1.4) 4 (1.9) .75a Statins 35 (5.2) 21 (9.0) .039 35 (5.6) 19 (8.9) .091 Class III antiarrhythmic drugs 9 (1.3) 234 (100.0) <.001 13 (2.1) 16 (7.5) <.001 Note: Categorical variables are presented as numbers (percentages). Abbreviations: AF, atrial fibrillation; AV, aortic valve. a Fisher's exact test.

Multivariate logistic regression models (model 2) showed that age (OR, 1.04; 95% CI, 1.01–1.08; P = .019), baseline left ventricular ejection fraction (OR, 0.97; 95% CI, 0.93–1.00; P = .027), new-onset AF (OR, 2.56; 95% CI, 1.15–5.70; P = .022) and use of beta-blockers after 3 months (OR, 0.45; 95% CI, 0.21–0.97; P = .041) were important predictors of persistent AF at 1 year (c-statistic = 0.754; Hosmer–Lemeshow test, P = .80; Table 4).

TABLE 4. Univariate and multivariate logistic regression models to identify predictors of persistent AF at 1 year Variables Univariate models Multivariate models OR (95% CI) P value OR (95% CI) P value Age per year 1.05 (1.01–1.09) .011 1.04 (1.01–1.08) .019 Male sex 1.48 (0.62–3.50) .38 Coronary artery disease 0.94 (0.22–4.05) .93 Diabetes 1.41 (0.42–4.80) .58 Hypertension 1.08 (0.51–2.28) .83 Chronic lung disease

留言 (0)

沒有登入
gif