AF and HF are considered cardiovascular “epidemics” of these days affecting millions of people. Both are chronic disorders, and patients are usually managed for years as out- or inpatients under the care of family physicians, internists, cardiologists, or subspecialists such as cardiac electrophysiologists or heart failure specialists. Even though the coexistence of AF and HF in the same patient is a common problem, the information on the clinical relevance of the order these entities present is limited.

In our study, the most common timing scenario was the simultaneous presentation of AF and HF, followed by HF → AF and the AF → HF sequences. Patients with simultaneous presentation included more women and patients from the oldest (80 +) cohort as compared with the two other groups. Although mixed results have been published in the few studies with similar data available (Table 2), a relatively even distribution was found with a significant representation of each timing scenario.

In the whole cohort including 100,004 patients with mortality data available Kaplan–Meier survival curves separated according to the sequence of AF and HF. Simultaneous presentation was associated with the lowest survival curve with early separation after 3 months. Curves related to the AF versus HF first sequences separated after 6 months indicating significantly better survival at a 5% difference in relative risk (RR) when HF followed AF, thereby confirming the findings in previous reports [16,17,18], including the recently published data from the Danish study [15].

To our knowledge, our study is the first one which analyzed mortality in different sequences in different age groups. The lowest survival was consistently demonstrated in simultaneously diagnosed patients in all age groups studied. In the 65–79-year range (Fig. 3b), which contained the largest cohort within the study, mortality rates in the different sequences followed the same pattern as in the whole cohort. In patients below 65 (Fig. 3a), representing the age group smallest in size, survival advantage of the AF → HF sequence was even larger as compared to the HF → AF and the simultaneous sequences (relative risk reduction [RRR] = 25% and 29%, respectively). As expected, better survival rates were detected in this youngest cohort with any sequence, thereby median values could not be calculated with 66 months follow-up. Strikingly low survival rates (4–7 months) were demonstrated in patients above age 80 with all scenarios (Fig. 3c). Furthermore, mortality rates with the AF → HF versus HF → AF sequences reversed in this age group: patients had better survival when AF presented in a previously diagnosed HF condition. However, differences with small absolute values mitigate the clinical relevance of the statistical significance of this comparison. Importantly, the endpoint of our study was all-cause mortality. Types of mortality (cardiovascular versus non-cardiovascular etc.) could not be captured from the database. Accordingly, some of the deaths in our cohort may not have related to either AF or HF.

We also analyzed mortality differences related to gender in the different timing scenarios and age groups. For the total patient population, better survivals were demonstrated in women with all three sequences. However, no sex difference in survival was demonstrated above age 80 with any patterns. The lack of survival benefit of female sex above 80 years is not surprising considering the dramatically low overall survival in this age group (Figs. 4, 5, 6).

Comparing data from the Danish [15] and the Hungarian databases, higher mortality rates were found in our patients at 1 year with both the AF → HF (35 vs. 25%) and the HF → AF (37% vs. 30%) sequences (Table 3). These differences became smaller at 3 years with these two sequences. However, dramatically lower survivals were detected in simultaneously diagnosed cases in our study at both follow-up timepoints.

We cannot provide a definite explanation for this difference between the results of two large-volume, nationwide European studies with many similarities in design. Although simultaneously diagnosed patients represented the oldest cohort in our, while the youngest one in Pallisgaard’s study, the mean and median ages (74.63 and 75 years) in the two cohorts were still comparable. However, life expectancies at birth in Denmark versus in Hungary were 79.4 versus 70.69 years for male patients and 83.1 versus 77.52 years for female patients, and mortality rates/1000 inhabitants were 9.84 versus 14.5, respectively (data from 2021) indicating significant differences in the general health status of the two populations. Although details of comorbidities in the two studies might have suggested at least a partial explanation for the differences, these data were not captured in our trial. The low median survival detected in our patients suggest that many of these individuals were probably latecomers, who were seeking medical help at an already advanced stage of their disease.

HF after AF was diagnosed after a shorter delay (median 6 months) than vice versa (median 10 months) in the whole patient population (Fig. 7). These time delays demonstrated slight changes with increasing age in both sequences (Figs. 8, 9). However, consistently shorter time delays were detected between the two comorbidities with the AF → HF scenario across all age groups.

The differences in both the time delays and in life expectancies demonstrated in our study may be related to the different underlying patomechanisms of the two scenarios. In the case when HF occurs first, the leading cause of developing AF is likely to be related to electromechanical remodeling, which starts later during disease progression, but indicates a more advanced stage of HF which is less likely to respond to therapy. In contrast, HF may develop earlier after AF as a result of tachycardia-induced left ventricular cardiomyopathy, the loss of atrial pump function, and the irregularity of the ventricular cycle lengths [19]. Patients with the latter scenario might respond better to evidence-based HF and antiarrhythmic therapy, especially when it is initiated early in the course of the disease. Arrhythmia-induced left ventricular dysfunction, referred to as tachycardiomyopathy may improve even with therapy targeting only the arrhythmia (rhythm or rate control) with partial or even full resolution of LV dysfunction. Furthermore, data from the CASTLE-AF Trial [20], and a few single-center studies [21] demonstrated that the restoration of sinus rhythm with left atrial ablation in HF patients may result in reverse remodeling of the LV and a reduction in all-cause mortality. Whether the order AF and HF present may modify the favorable effect of catheter ablation in these patients remains to be clarified.

Among other studies including patients with concomitant AF and HF, our study enrolled the largest number of patients. We confirmed previous findings in the literature in that the sequence of presentation of these conditions in the same patient have significant prognostic implications. Furthermore, the survival benefit with the AF → HF sequence is larger at younger age, while the differences in magnitude are shrinking with age, especially above 80 years. The survival advantage of female sex also disappears in the oldest cohort. These observations suggest that family physicians, internists or cardiologists should pay attention to the sequence these conditions present. Newly discovered AF in a stable patient with HF should be considered as a “red flag” of possibly rapid deterioration. The closer monitoring and early intensification of available therapy to the maximum level is justified involving heart failure and arrhythmia specialists in these cases. Although recent developments in the pharmacotherapy for HF including the novel classes of angiotensin receptor-neprilysin inhibitor (ARNI) and sodium-glucose cotransporter-2 (SGLT2) inhibitors are based on the results of large-scale multicenter trials, no results are available on those patients who had coexisting AF and HF with different orders of presentation. Evaluating the array of therapeutic options in this specific cohort should therefore be an important consideration while planning future studies. In addition, the sequence AF and HF present should be considered while assessing the benefit of pharmacotherapy and catheter ablation strategies.

This is a longitudinal database study which included more than 100,000 patients. Major limitations of our investigation are that no data were collected on comorbidities and on treatment of these patients. Study participants were identified by ICD codes with no details on the types of AF and HF and the exact cause of death remained unknown. Novel HF medication options like ARNI and SGLT-2 inhibitors were not available during the study period. However, the large number of patients, including in- and out-patients collected from a database covering the whole Hungarian population, and the more than 5-year follow-up adds strength and significance to our findings.