41 SEX- AND AGE SPECIFIC ASSOCIATION OF AF WITH MORTALITY IN COVID-19 tract infections (e.g. influenza). Nevertheless, for full appreciation the following topics deserve attention. First, while our findings show that AF and/or AFL appears prognostically unfavourable, particularly in males, this does not imply a causal relationship. In fact, it could be argued that the development of AF/AFL and its impact on mortality is merely a more general signal of progression of disease severity and accumulation of comorbidities (e.g. exemplified by higher CHA2DS2-VASc scores), and thus could be considered as an ‘innocent bystander’ in patients experiencing clinical deterioration. To explore the impact of the development of new-onset AF and/or AFL during hospitalisation on in-hospital mortality when adjusting for concomitant comorbidities and risk factors, we performed a sensitivity analysis with additional adjustment for CHA2DS2-VASc score. Although this analysis is inherently impacted by a lower degree of statistical robustness due to missing information on the CHA2DS2-VASc score in 75.7% of patients (n=2,779), it did yielded similar inferences (Supplementary File S3 and S4). Moreover, in our study the majority of AF and/or AFL cases (60.7%) were detected in patients either before ICU admission or in patients never admitted to the ICU (i.e. before widespread increase in disease severity occurred). Although it is possible that the threshold for ICU referral was higher due to limited capacity during the peak of the pandemic, this suggests that (new-onset) AF and/or AFL, would at least be an early marker for disease progression. Based on our findings this appears to be prognostically unfavourable, particularly in males aged between 60 and 72 years. Second, although diagnoses were centrally defined, with multicentre studies there is always a risk of heterogeneity due to differences in interpretation among centres. Given that the strategy for rhythm monitoring was defined by the attending physicians, and as a consequence was different per centre, it could well be that electrical disorders may have been underdiagnosed in patients on general wards where continuous rhythm monitoring is not performed. Moreover, grouping the different AF subtypes (paroxysmal, persistent, and permanent) may have resulted in missing subtle disease progression within the AF group. Finally, since only in-hospital death could be recorded, mortality outcome data are limited, and comparison with other studies is hampered by differential follow-up due to differences in length of hospital stay. CONCLUSION Using a large international database, this study confirms that AF and/or AFL is the most prevalent electrical disorder in hospitalised COVID-19 patients, and that new-onset AF and/or AFL is associated with a poorer prognosis exemplified by an increased inhospital mortality. However, this increased mortality risk appears to be restricted to male patients aged between 60 and 72 years, and was not observed in female patients. 3
RkJQdWJsaXNoZXIy MTk4NDMw