In our single center registry study, no significant difference could be seen regarding the occurrence of ventricular arrhythmias and ICD therapies in ICD patients with remote patient monitoring comparing two time intervals during and one time interval after the first national lockdown in Germany due to the COVID-19 pandemic and a time interval 1 year prior to the COVID-19 pandemic. There was also no difference in respect to clinical events, which were very low during each time interval. During the COVID-19 pandemic, there were significantly fewer planned visits to our ICD outpatient clinic, which did not result in a worsening of the rhythmologic or clinical condition of ICD patients with remote patient monitoring.
In-person-visits to the ICD outpatient clinic
When the COVID-19 pandemic reached Europe and the United States in March 2020 in-person-visits to outpatient providers declined. In our study, in-person-visits to our outpatient ICD clinic were almost halved during the first national lockdown in Germany compared to the same time interval 1 year before. Planned visits even declined by 89%, whereas unplanned visits showed no difference. There were still significantly fewer planned visits during the 7 weeks after the first lockdown and during the second lockdown. On one hand hospitals were still trying to reduce in-person visits, on the other hand patients themselves also were trying to avoid visits to the hospital to avoid a possible risk of infection. In an analysis of US insurance data from January 2020 to June 2020 a decline of in-person-visits to outpatient providers by 30.0% could be seen [12]. Another study that presented data from the US Outpatient Influenza-like Illness Surveillance Network saw a 70% decline of outpatient visits during April 5th to April 11th, 2020 compared to the same period 1 year before [13]. In the current study, the strong decline in planned visits was due to the fact that only patients with remote patient monitoring were investigated.
Ventricular arrhythmias and ICD therapy
The current study showed no significant difference between any time interval regarding the occurrence of ventricular arrhythmias and ICD therapies. Similar results were found by a recent study by Sassone et al. [14] The authors compared a 10-week long time interval of the first lockdown in Italy to the time interval before the lockdown began and the corresponding time interval 1 year earlier. No differences in the occurrence of arrhythmias and ICD therapies were found in this study. In difference to the current study, the investigators included all patients with an ICD regardless of whether the patients had access to remote patient monitoring. Furthermore, the time interval after the lockdown and the time interval during the second lockdown were not investigated. Another current study compared ICD patients in whom the regular in-person-visit was replaced by a remote patient monitoring interrogation to patients who recently had their regular in-person-visit. Regarding the occurrence of arrhythmic events, no significant difference between the two groups could be found [15]. The study design of this study was completely different compared to the current study. The authors compared an intervention group of 131 patients to a control group of 198 patients during 1 month of the pandemic, whereas the present study investigated the differences of one group of 140 ICD patients with remote patient monitoring during 4 different 7-week long-time intervals during and before the pandemic. Therefore, it was conducted over a longer period and also investigated the effect of different time intervals of the pandemic with different lockdown measures on the rhythmologic and clinical situation of ICD patients with remote patient monitoring. Several studies have shown that an infection with COVID-19 could be a potential trigger for arrhythmias [16,17,18]. Although atrial arrhythmias were detected more frequently in patients hospitalized with COVID-19, ventricular arrhythmias may especially play a role in patients with preexisting cardiac conditions, such as ischemic heart disease [17, 19]. A case report even described a patient with an electrical storm, which was potentially triggered by an infection with COVID-19 [20]. Several mechanisms have been discussed to explain the high incidence of arrhythmias in patients hospitalized with COVID-19. Cytokine storm, mediated by an imbalanced response among subtypes of T-helper cells and hypoxia-induced intracellular calcium overload leading to early afterdepolarization can contribute to trigger ventricular arrhythmias [19,20,21,22]. Myocardial injury, myocarditis, direct viral invasion or the use of QT prolonging drugs such as hydroxychloroquine and azithromycin may also lead to the occurrence of ventricular arrhythmias [18, 23,24,25,26]. On the other hand, a study by Gasperetti et. al that described ECG modifications and arrhythmic events in COVID-19 patients undergoing hydroxychloroquine therapy found only modest QTc prolongation, a low ventricular arrhythmia rate of 1.1% and no arrhythmic-related deaths [27]. In the current study, only one patient had an infection with COVID-19 during LD1. This patient was not hospitalized or treated with hydroxychloroquine or azithromycin. No further infections occurred during the other investigated time intervals of the pandemic. Therefore, we can exclude that the beforementioned factors had a relevant influence on the number of ventricular arrhythmias in the present study. The reason for this finding may be that the lockdown measures during the first year of the pandemic and the behavior of chronically ill patients with an ICD, who may have been very careful to avoid an infection, could have led to a very low incidence of COVID-19 infections in the study group.
Clinical events
It's still a matter of debate, whether the decline in patient visits results in a worsening of chronic illnesses, such as chronic heart failure or coronary artery disease. Clinical events were low during every time interval, and no difference could be seen to the time interval 1 year prior to the pandemic. Only one patient died between the time intervals postLD1 and LD2. Comparable to the present findings, the aforementioned study by Sassone et al. also found no increase in arrhythmic death during the first national lockdown in Italy [14].
Unplanned visits to our ICD outpatient clinic, including problems detected by remote patient monitoring, hospitalized patients and patients presenting in the emergency room showed no significant difference during the different time intervals of the pandemic.
As the investigated time intervals were only 7 weeks long, one can still speculate whether changes in patients’ behavior during the pandemic, such as reduction of physical activity and social distancing, might have any influence on the medical condition of chronically ill patients over a longer time period.
In contrast to the present findings, a nationwide survey in Italy found a reduction in admissions for acute myocardial infarction of 48.4% comparing a week in March 2020 to the equivalent week in 2019, but this resulted in a substantially increased STEMI fatality rate [28]. As this was a nationwide survey, which looked at the outcome of patients with acute myocardial infarction, the results cannot be compared to our single center study with a completely different group of patients. ICD patients often have a long history of chronic heart failure or coronary artery disease and might recognize an increase of symptoms early enough to contact their medical provider and avoid further worsening.
A rise of the arrhythmic burden could be an early indicator for a worsening of chronic cardiac conditions. We didn't see any difference regarding the number of patients with ventricular arrhythmias during any of the three time intervals of the pandemic and the interval before the pandemic. We can just speculate what the possible reasons might be, but the benefit of remote patient monitoring, which already has been described in pre-pandemic times, might also play an important role during the pandemic. The ALTITUDE Survival Study found a 50% reduction of mortality in patients with ICD and CRT-D and remote patient monitoring compared to usual care patients without remote patient monitoring [29]. Other studies have shown that remote patient monitoring enables the physician to see an episode of an arrhythmia faster and to react for example with further clinical examinations or adjustment of the medical therapy [9, 10, 30, 31]. Furthermore, patients who accept remote patient monitoring might be more compliant and more aware of their general health status and might seek a medical consultant early enough to avoid a further deterioration of their chronic condition. In sum, the close surveillance of the rhythmologic situation and the good compliance of ICD patients with remote patient monitoring could be possible reasons, why these patients are less negatively affected by a reduction of planned in-patient-visits to their medical provider.
The TRUST trial found that in a group of patients with an ICD and remote patient monitoring, the in-clinic and hospital visits could be reduced by 45% compared to the conventional care group. Despite fewer hospital visits, the detection of device specific events was advanced by 30 days and no difference in clinical adverse events could be seen [9]. Although the TRUST trial was conducted before the pandemic, the findings are very similar to the current study. Only one patient died between postLD1 and LD2, which may show that outpatient clinic visits for device interrogation can be reduced safely for ICD patients with remote patient monitoring during times of the pandemic without an increase of mortality and other clinically relevant events. Finally, the reduction of in-person-visits is an important aspect during a global pandemic regarding the reduction of potentially contagious situations for the patient and medical staff.
Limitations
As we only included patients with remote patient monitoring and had no control group of ICD patients without remote patient monitoring, no statement can be made about the safety and outcome of ICD patients in general. Each investigated time interval was 7 weeks long due to the length of the first national lockdown in Germany. Therefore, we cannot say, whether a longer duration of lockdown measures would have caused other results. The retrospective non-randomized design makes a selection bias possible, as patients with remote patient monitoring are potentially more compliant and seek medical treatment early enough to avoid clinically relevant endpoints. Furthermore, we just registered hospitalizations in our institution. Therefore, we do not know whether patients were hospitalized in external hospitals. Because of the study design, we don’t have information about the number of COVID-19 infections in the study cohort during the different time intervals of the study. Finally, this was a single center study, and we do not know if our results can be generalized to other medical centers that might have other strategies for selecting patients eligible for remote patient monitoring.