Incidence of and causes for all-cause hospitalizations in patients with atrial fibrillation

Background: Atrial fibrillation (AF) is the most common cardiac arrhythmia in clinical practice, and its prevalence is expected to further increase in the future. AF patients not only have a high number of comorbidities, but they also have an increased risk of hospital admissions compared to individuals without AF. Nevertheless, predicting hospital admission risk among patients with AF remains difficult, and possible preventive strategies unclear.
Based on these gaps in knowledge, the overall goal of this PhD thesis was to investigate the incidence of and causes for all-cause hospital admission in patients with AF. The specific aims were (1) to perform a systematic review and meta-analysis summarizing the current evidence of the incidence of and associated risk factors for hospital admissions in AF patients; (2) to identify risk factors for hospital admissions in our own cohorts and subsequently use this knowledge to develop and validate a risk score for predicting hospital admissions; (3) to identify psychosocial factors associated with hospital admissions in patients with AF.
Methods: For the meta-analysis, we performed a comprehensive literature search in PubMed, EMBASE and CENTRAL, and pooled incidence rates for hospital admissions using random-effects models. Factors associated with observed between-study heterogeneity were identified using meta-regression analysis.
For the second and third study, we used data of two ongoing, prospective observational cohort studies, the Basel Atrial Fibrillation Cohort Study (BEAT-AF) and the Swiss Atrial Fibrillation Cohort Study (Swiss-AF) in which 3,968 patients with diagnosed AF were enrolled. Unplanned hospital admissions were defined as any unpredicted admission leading to at least one overnight stay. For the second study, we used the Swiss-AF data set as the derivation cohort and performed a variable selection using the least absolute shrinkage and selection operator (LASSO) method. Multivariable adjusted Cox regression analyses were performed to assess the effect of the selected variables on all-cause hospitalization. Based on regression coefficients we constructed a risk score and subsequently validated the score in the external validation cohort (BEAT-AF).
For the third study, we used psychosocial factors, such as marital status, education, level of depression and health perception, and investigated their effects on risk of hospital admission. Cox regression analyses adjusted for conventional risk factors for hospital admission were performed to calculate hazard ratio (HR).
Results: We included 35 studies of 311’314 AF patients in the meta-analysis. The pooled incidence of all-cause hospital admissions was 43.7 per 100 person-years. AF patients were more often admitted for cardiovascular causes (26.3 per 100 person-years), but the risk of non-cardiovascular hospitalizations was substantial (15.7 per 100 person-years). Associated factors for hospital admission were older age, longer follow-up time and prevalent chronic pulmonary disease or cancer.
In the second study we found that the most important predictors for all-cause hospital admission were age (75-79 years: adjusted hazard ratio [aHR], 1.33; 95% confidence interval [95% CI], 1.00-1.77; 80-84 years: aHR, 1.51; 95% CI, 1.12-2.03; 85 years: aHR, 1.88; 95% CI, 1.35-2.61), prior pulmonary vein isolation (aHR, 0.74; 95% CI, 0.60-0.90), hypertension (aHR, 1.16; 95% CI, 0.99-1.36), diabetes (aHR, 1.38; 95% CI, 1.17-1.62), coronary heart disease (aHR, 1.18; 95% CI, 1.02-1.37), prior stroke/TIA (aHR, 1.28; 95% CI, 1.10-1.50), heart failure (aHR, 1.21; 95% CI, 1.04-1.41), peripheral artery disease (aHR, 1.31; 95% CI, 1.06-1.63), cancer (aHR, 1.33; 95% CI, 1.13-1.57), renal failure (aHR, 1.18, 95% CI, 1.01-1.38), and previous falls (aHR, 1.44; 95% CI, 1.16-1.78). A risk score with these variables was well calibrated, and achieved a C statistic of 0.64 (95% CI, 0.61-0.66) in the derivation and 0.59 (95% CI, 0.56-0.63) in the external validation cohort.
In the third study including patients from Swiss-AF, 1582 (67.1%) were married, 156 (6.6%) were single, 287 (12.2%) were divorced, and 333 (14.1%) were widowed. Two hundred and seventy six patients (11.7%) had at most a primary education, 1171 (49.7) had secondary education, and 911 (38.6%) had a college or university degree. Depression or depressive symptoms was present in 99 (4.2%) patients. Median health perception was 75 (interquartile range [IQR], 60-85) on a scale ranging from 0-100, with higher scores indicated better health perception. The highest risk of all-cause hospital admission was observed in single (aHR, 1.35; 95% CI, 1.05-1.75) or divorced patients (aHR, 1.26; 95% CI, 1.03-1.54), and in those who reported low health perception (aHR for <75 points, 1.40; 95% CI, 1.21-1.61).
Conclusions: The overall incidence of hospital admissions in patients with AF is high. The risk of hospital admissions is related to multiple cardiovascular and non-cardiovascular risk factors, including several psychosocial factors and subjective health perception.
Outlook: Given the high risk among AF patients of being admitted to the hospital and the high burden of associated risk factors, new multidisciplinary preventive strategies are needed with the goal to reduce hospital admissions, unfavorable patient outcomes and healthcare costs.