40 CHAPTER 3 Sex- and age-dependent effect on AF on mortality Our study confirms previous (smaller) studies which reported AF/AFL as the most prevalent arrhythmia in COVID-19 patients, in addition to its association with increased mortality. With respect to AF/AFL occurrence, Peltzer et al. and Mountantonakis et al. observed a slightly higher prevalence of AF/AFL compared to our study (16% and 18% compared to 12% respectively), whereas Musikantow et al. found a similar prevalence of 10%.8,9,20 Conversely, Bhatla et al. reported a much lower prevalence of new-onset AF (3.5%), yet in a much smaller dataset of 700 patients.10 Similar to our study, Peltzer et al. and Mountantonakis et al. found AF and/or AFL, as well as new-onset AF and/or AFL, to be associated with increased in-hospital mortality.8,9 Bhatla et al. did not find such an association between new-onset AF and/or AFL and in-hospital mortality, yet (again) this study included a relatively small dataset with only 25 incident AF cases reported.10 Importantly, using the latest prediction methodology (allowing age to remain continuous in all analyses using cubic spline functions), we were – for the first time – able to pinpoint the effect of AF/AFL occurrence on in-hospital mortality to male hospitalised COVID-19 patients aged 60–72 years. In fact, we ruled out an effect of AF/ AFL occurrence on mortality in female patients with COVID-19, while in the general population females with AF/AFL have a worse outcome compared to males.12 As an example, in a male hospitalised COVID-19 patient of 65 years, the occurrence of AF/AFL would increase his risk of mortality from ±15% to ±35%, whereas in a female patient of 65 years this risk remains well below ±15–20%, regardless of AF/AFL development. More importantly, the correlation between age and its interaction with AF and/or AFL follows a non-linear pattern, which is even different for males and females, thus underlining the importance of our statistical approach. As such, our analyses provide a much more granular assessment of the effect of new-onset AF and/or AFL during hospital admission for COVID-19 by better identifying subgroups of patients where the prognostic impact on mortality is most relevant. Strengths and limitations Major strengths of our work include the inclusion of a large international dataset of nearly 6,000 hospitalised COVID-19 patients, allowing to perform sophisticated analyses on the incremental impact of AF/AFL occurrence on in-hospital mortality beyond the effects of age and sex. However, our findings might not be restricted to or typical for COVID-19 patients. For example, a recent study by Musikantow et al. shows a similar increase in mortality in hospitalised influenza patients with AF/AFL.20 This seems to indicate that the found association might be related to a general viralinduced systemic illness rather than specifically COVID-19, suggesting that the findings in this study might be generalised to other patients with viral induced respiratory
RkJQdWJsaXNoZXIy MTk4NDMw