Here, we present a German nationwide real-world analysis of 40,435 hospital cases in which an LAAO procedure was performed. Three main findings stand out: (1) The incidence of LAAO procedures in men is more than twice as high as in women, (2) there is a relevant shift towards an older patient collective, and (3) a higher age is associated with a longer hospital stay, more severe complications and a higher in-hospital mortality.

The database used in this study had been investigated two times before with respect to LAAO procedures: Hobohm et al. examined all cases in which the previous OPS-Code for LAAO procedures was used (2011–2015; 15,895 patients) [9]; Maier et al. reported data from 2016 to 2020 (28,039 patients) and focused on the difference between endocardial occlusion and epicardial loop stitch [10]. Our study covers an additional 12,000 patients and an integration with data of the general German population, both of which allowed for a more in-depth analysis of gender and age trends and their impact on intrahospital outcome in patients undergoing LAAO.

Regarding gender distribution, our analysis showed a clear surplus of treated men compared to women in all age groups. This difference seems least pronounced in the highest age group (46.7% female) but in relation to the gender distribution of the total German population, the overall incidence of LAAO in women calculates to be only half as high as in men of the same age. This is a striking finding, given the reported lifetime risk of atrial fibrillation being just slightly higher in men [16] and given the evidence that women are found to potentially have more bleeding complications under OAC [17]. On the other hand, it is in line with data from the US, where in a large nationwide cohort the proportion of women was also significantly lower than the gender distribution would imply [7]. A surplus of men is a well-known phenomenon in interventional cardiology and applies to several indications [18]. We see the same and well-described biases in charge also for the field of LAAO as also in the initial randomized-controlled trials women were underrepresented [4]. The fundamental question of whether interventional techniques initially applied mainly in male patients can be transferred and equally applied in women is still open. For LAAO there is a signal from a meta-analysis of 16 studies (111,775 patients) that the periprocedural complication rate is higher in women [8]. In our analysis, female gender was also associated with a higher risk of in-hospital death, but this was not statistically significant.

Our finding of a clear shift towards an ever-aging patient population was not discussed in previous publications, but the analysis of Hobohm et al. had already shown an increasing mean age over the investigated time period. From 2011 to 2022 the median age rose from 75 (IQR: 70–79) to 79 (IQR: 72–83) years. While patients in the 75–80 age group still accounted for the majority of patients treated in 2016, patients in the 80–85 age group made up the majority in 2022. Two US-American nationwide sample studies on patients undergoing LAAO have been published [7, 19]. Median age is comparable to our population. Both do not report age trends for the much shorter time periods analyzed. Freeman et al., however, report similar age categories with the group with the highest age already comprising a high number of patients (14.1%) in the years from 2016 to 2018.

In line with the two previous nationwide German analyses, we also found no significant improvement in procedural safety over the investigated years (in-hospital mortality ranged from 0.9 to 1.4% from 2013 to 2020 [10]). A recently published analysis from a prospective multicenter German registry [20], including 638 patients from 38 hospitals between 2014 and 2016, reported significantly fewer MACE events (0.6%) and fewer combined major complications (4.4%). It should be noted here that the patient collective was on average 4 years younger than in the total German cohort reported here. Compared to the US NCDR’s LAAO Registry 2016–2018, German patients were slightly younger and had a higher percentage of permanent atrial fibrillation and a higher incidence of severe comorbidities like chronic kidney disease (38.2% vs 13.6%). Notably, the in-hospital mortality was more than fivefold higher (1.1% vs. 0.19%) in the German cohort. This is a striking finding, also as in-hospital mortality in the other US-American nationwide study was similarly low at 0.14% [19]. Here the analyzed time period and patient age again were very comparable to the NCDR’s Registry, differentiated information on comorbidities was sparse though. Further real-world data from the US also took readmissions within 90 days into account and found a mortality rate of 0.53% [21]. Even though this is more than double the rate of the reported US in-hospital mortality in the nationwide samples, it still is lower than in Germany. Most recent evidence from an all-comers registry, the Amulet registry (Europe, Australia, Israel, Chile, Hong Kong), showed 3 procedure-related deaths (0.28%), further mortality rates were 0.59% within 30 days 8.4% within the first year [22]. In all four randomized controlled trials analyzing LAAO only one procedure-related death (0.5%) was reported in the PRAGUE-17 study, in all others there was none [2,3,4, 23].

The differences in reported mortality rates between Germany and the US nationwide cohorts are difficult to understand. Due to the nature of retrospective cohort studies more granular data on timing and etiology of mortality is not available. In general, interventional technique and periprocedural care should by all means be comparable in the US and Europe. Nonetheless, we also noted higher rates of severe complications that may potentially have led to more intra-hospital deaths (i.e., the rate of cardiac surgery in our population was 2.0% compared to 0.24% in the NCDR’s LAAO Registry). Further, we consider two differences in patient management that may explain the different outcomes. One, as eluded to before, the patients we report appear to be significantly sicker. Secondly, it may be for a different timing of indication and intervention. The more LAAO is undergone as an elective procedure the less comorbidities should play into role. Irrespectively, a reporting bias cannot be ruled out, we believe though that mortality as the most definite outcome should be explicitly well coded in Germany and the US.

How relevant patient characteristics are for the occurrence of MACE is shown by our analysis as rates of in-hospital mortality, pericardial effusion, acute kidney injury and even heart surgery are highest in the group with the highest patient age (in-hospital mortality 2.2%, MACE 7.4%). Balancing the risks and benefits of an interventional procedure is crucial especially in a prophylactic treatment like LAAO. With increasing age, the risk/benefit ratio seems to be negatively impacted as already in-hospital MACE rates are higher than predicted annular stroke rates (see CHA2DS2VASc-Score) [7]. Notably, also transfusion rates were highest in the oldest age group. It is unclear whether these were triggered by bleeding complications or reflect the presumed indication leading to LAAO—a relevant bleeding event under OAC.

Definitely, these data underscore the need for prospective randomized trials addressing patient populations that differ from the initially investigated one by age and also by gender to provide further evidence for the pros and cons of interventional stroke prevention in patients with atrial fibrillation.

This study is based on a DRG registry containing ICD and OPS discharge codes coded primarily for reimbursement. This could lead to under-and overreporting of well-reimbursed procedures and diseases. The endpoints presented here were selected against this background; we are convinced that reporting biases are equally distributed in the groups presented. On the other hand, we present real-world data without selection bias as in RCTs and reflect the reality of care in Germany.