Utilization and in‐hospital complications of catheter ablation for atrial fibrillation in patients with obesity and morbid obesity

Abstract Background Real‐world data on atrial fibrillation (AF) ablation outcomes in obese populations have remained scarce, especially the relationship between obesity and in‐hospital AF ablation outcome. Hypothesis Obesity is associated with higher complication rates and higher admission trend for AF ablation. Methods We drew data from the US National Inpatient Sample to identify patients who underwent AF ablation between 2005 and 2018. Sociodemographic and patients' characteristics data were collected, and the trend, incidence of catheter ablation complications and mortality were analyzed, and further stratified by obesity classification. Results A total of 153 429 patients who were hospitalized for AF ablation were estimated. Among these, 11 876 obese patients (95% confidence interval [CI]: 11 422–12 330) and 10 635 morbid obese patients (95% CI: 10 200–11 069) were observed. There was a substantial uptrend admission, up to fivefold, for AF ablation in all obese patients from 2005 to 2018 (p < .001). Morbidly obese patients were statistically younger, while coexisting comorbidities were substantially higher than both obese and nonobese patients (p < .01) Both obesity and morbid obesity were significantly associated with an increased risk of total bleeding, and vascular complications (p < .05). Only morbid obesity was significantly associated with an increased risk of ablation‐related complications, total infection, and pulmonary complications (p < .01). No difference in‐hospital mortality was observed among obese, morbidly obese, and nonobese patients. Conclusion Our study observed an uptrend in the admission of obese patients undergoing AF ablation from 2005 through 2018. Obesity was associated with higher ablation‐related complications, particularly those who were morbidly obese.

younger, while coexisting comorbidities were substantially higher than both obese and nonobese patients (p < .01) Both obesity and morbid obesity were significantly associated with an increased risk of total bleeding, and vascular complications (p < .05). Only morbid obesity was significantly associated with an increased risk of ablation-related complications, total infection, and pulmonary complications (p < .01). No difference in-hospital mortality was observed among obese, morbidly obese, and nonobese patients.

| INTRODUCTION
Obesity is a well-established risk factor of both new-onset and recurrent atrial fibrillation (AF). 1 In obese patients, Several mechanisms promote AF, not only coexisting comorbidities but also inflammatory processes and atrial electrophysiological remodeling. For these reasons, American College of Cardiology/American Heart Association guidelines have recommended weight loss by comprehensive weight management programs to mitigate AF burdens and severities. 2,3 On the contrary, several studies have pointed out the obesity paradox effect owing to the favorable cardiovascular outcomes observed on anticoagulated AF patients with obesity compared to nonobese. 4 Furthermore, one study suggested weight loss in obese patients who underwent AF ablation, surprisingly, did not improve AF recurrence rates. 5 Catheter ablation has been widely accepted as a principal treatment for AF 6 and has become the first line in particular subgroups such as younger and athletic patients. 7 The majority of ablation outcomes in large clinical trials have come from exclusively high-volume academic centers with experienced operators. As AF catheter ablation has been performed broadly, the results can be varying across regions depending on center experiences and practices. Real-world data on AF ablation outcomes in obese populations have so far been scarce, especially the relationship between obesity and in-hospital AF ablation trend and outcomes.
Thus, we evaluated the most recent admission trend and the impact of obesity in patients undergoing AF ablation on the hospital outcomes by utilizing the National Inpatient Sample (NIS) database, the current largest inpatient database from US inpatient settings.

| Study patients and variables
International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) was used for reporting diagnoses and procedures in the NIS database during the study period. For each index hospitalization, the database provides a principal discharge diagnosis, additional diagnoses, and procedures, in which total maximum numbers vary per year. We identified patients 18 years of age or older with a primary diagnosis of AF (ICD-9-CM 427.31 and ICD- To determine the impact of obesity in patients undergoing AF ablation, body mass index (BMI) was used to stratify patients into nonobese (<30 kg/m 2 ), obese (≥30 and ≤40 kg/m 2 ), and morbidly obese (>40 kg/m 2 ). 8  The following patient demographics were collected from the database, including age, sex, and race. In addition, associated comorbidities were identified by measures from the Agency for Healthcare Research and Quality. For the purposes of calculating Deyo-Charlson Comorbidity Index (Deyo-CCI), an additional list of comorbidities was identified from the database using ICD-9-CM codes and ICD-10-CM codes (Table S1). Deyo-CCI is a modification of the CCI, containing 17 comorbid conditions. Higher Deyo-CCI indicates a more severe condition and is an indicator of patient mortality 1 year after an admission.

| Outcomes
We identified the common in-hospital complications of CAs using the ICD-9-CM and ICD-10-CM diagnosis, and procedure codes using the same methodology as described in previous publications. 9 Table S2.
Readmissions for any reasons and outcomes which occurred after discharge were unavailable for the analysis as these data were not included in the NIS data set package.

| Statistical analysis
Trend weight files provided by Agency for Healthcare Research and Quality were used to reflect national estimates. The χ 2 test and analysis of variance were used to compare categorical and continuous variables, respectively. Trends for continuous variables were tested using the nonparametric test for trend by Cuzick. 13 In compliance with the data use agreement of HCUP nationwide databases, it was recommended to avoid reporting small numbers of observations (≤10) to minimize risks of person identifications.
To account for hospital-level clustering of discharges, we generated a two-level mixed-effects logistic regression model to identify independent predictors of complications. 9  with Medicaid or Private as a primary payer. Hypertension, congestive heart failure, diabetes, and chronic pulmonary disease were the most prevalent comorbidities. AF ablation was mostly performed in teaching hospitals (Table 1).
Coexisting comorbidities were found significantly higher among all obese patients when compared to nonobese patients. The highest proportion of patients with Deyo-CCI ≥ 2 was observed in morbidly obese patients, followed by obese and nonobese patients (p ≤ .001).
The use of anticoagulation was highest in morbidly obese patients, followed by obese and nonobese patients (p ≤ .001; Table 1).  Figure 1A). Essentially, the admission rates for AF ablation in morbidly obese patients were substantially rising from 1.5% to 21.3% (p < .001; Figure 1B). On the other hand, a significant downtrend was found from 2005 to 2018 among nonobese patients, potentially by a shift into an outpatient-based procedure (p < .001), while a nonstatistical decrement in the admission trend of obese patients was observed (p = .114; Figure 1C).

| Influence of obesity on in-hospital complications and deaths
For ablation-related complications, it was found 10.4% of the total cohort. Of these, total bleeding rates were the most common complication (4.9%), followed by total infection rates (2.8%) and total pericardial complications (1.7%). By obesity classifications, morbidly obese patients had the highest rates 14.3%, followed by obese patients 11.4% (p ≤ .001;  Figure 3 and Table S5).
Of note, neither obesity nor morbid obesity was associated with increases in total cardiac complications, total GI complications, total vascular complications, total skin complications, total neurological complications, and total pericardial complications.
In-hospital mortality rates in our cohort were extremely low, only 0.2%. There was no statistical difference in mortality rates between nonobese, obese, and morbidly obese patients (p > .05). F I G U R E 2 In-hospital complications trend in all obese patients who underwent AF ablation. AF, atrial fibrillation obesity and morbid obesity were associated with total vascular complications and bleeding events. In particular, only morbid obesity was statistically correlated with higher ablation-related complications, infection, and pulmonary complications, while only a trend toward higher ablation related complications was found in obese patients Nevertheless, both obesity and morbid obesity were not associated with an increased risk of total cardiac, pericardial, GI, skin, and neurological complications. Finally, total hospital mortality rates were only 0.2% in total of our cohort, which was no different among obese, morbidly obese, and nonobese patients.

| Obesity and its in-hospital complications impact
Very few studies described the potential complication risks of being overweight. One study 18 found morbid obesity had up to 3.1-fold complication rates, and 5% per 1 unit increment in BMI, while the other 19 only showed a modest increase in total complications.
However, the preponderance of previous studies negated the influence of obesity on complications. [14][15][16] Of these, all obese patients were gathered into one group as one variable, which may confound its true association. Also, complication rates were not the outcome of interest, precluding the complete investigation. Contrastingly, our objective was to resolve this conflicting data, explicating whether obesity portends any adverse outcomes in the setting of hospitalization for AF ablation. Conclusively, both obesity and morbid obesity were associated with higher complication rates compared to nonobese patients. This finding emphasizes the disadvantage of obesity, despite the "obesity paradox" previously described in anecdotal literature. 20,21 In obese patients with BMI 30-39, total complication rates were implicated by a substantial portion of vascular and bleeding complications. This finding was intriguingly opposite to a myriad of former studies which described favorable outcomes in the obese population. [22][23][24] In those studies, thinner patients were older and frailer and had more comorbidities, of which all were associated with poorer outcomes. Moreover, the bleeding incidence was lower among patients with higher BMI undergoing cardiac intervention, 25 To the best of our knowledge, no prior studies were examined whether obesity was related to infection risks in cardiac intervention settings, aka PCI or catheter ablation. Our study is the first to show this paramount correlation.

| Limitation
Our study has several limitations. First, as the NIS database provided only inpatient information, readmissions for any reason were unanalyzable owing to unprovided information. Also, residual biases from retrospective design cannot be excluded, despite a comprehensive covariate adjustment in our analysis. Second, owing to the same reason, this study may over/underestimate the true impact of obesity given the recent AF ablation trend in an outpatient setting.
Furthermore, susceptibility to error coding is hardly evitable in administrative databases. Despite this limitation, this is the largest representative data of an inpatient AF ablation best reflecting realworld experience. Third, as BMI was used to stratify obesity classifications, it may not truly represent adiposity in certain populations, particularly the younger population, which correlated more to lean body/muscle mass. 41 Obese status in the current databases may miscoded given this possible issue. Fourth, some factors which may affect the outcomes, for example, medications, especially types of anticoagulation, procedural techniques, fluoroscopy time, and use of hemostasis device, are not provided by NIS. Fifth, delayed onset complications, such as pulmonary vein stenosis and atrioesophageal fistula, cannot be accurately assessed in this study as these complications tend to occur after discharges. Moreover, the efficacy of AF ablation was unable to be determined owing to no available ICD codes for this aspect.

| CONCLUSION
Based on the largest all-payer inpatient database in the United States, our study observed an increase in trends of complications in obese patients undergoing AF ablation. Moreover, obesity was associated with higher ablation-related complications, particularly those who were considered morbidly obese.

DATA AVAILABILITY STATEMENT
The data for this systematic review and all potentially eligible studies are publicly available through the Open Science Framework (https:// osf.io/p7d6v/).