Within a very large, contemporary study analyzing nearly 123,000 long-term continuous ambulatory cardiac monitors that were prescribed between 2011 and 2013, we found a moderate burden of potentially high-risk arrhythmias, including both ventricular arrhythmias and bradyarrhythmias. Patient compliance with extended monitoring was high, with at least 25 % of patients achieving greater than 13 days of continuous monitoring. For all arrhythmias examined, longer monitoring times significantly increased the yield of detected arrhythmias. While the gains in arrhythmia yield were particularly marked in the first 7 days of monitoring, it is notable that the gains continued to increase from days 7 to 14. The relatively high wear and analyzable time for the longer term continuous monitors suggests that outpatient ECG monitoring using this approach is feasible and can have significant yield of clinically important arrhythmias beyond atrial fibrillation. Our study examined a more recent time period than prior investigations of long-term continuous monitors [5, 9] and suggests that changes in the device technology and accumulated operator experience may have resulted in improved patient compliance.

Prior research suggests that traditional 24-h Holter monitoring is not sufficiently long enough to detect many types of arrhythmias [10–13], and recent evidence has demonstrated that longer monitoring may be useful to detect arrhythmias in high-risk patient populations, such as those with a recent history of cryptogenic ischemic stroke [8, 14], although the majority of these studies have focused primarily on finding atrial fibrillation. For example, in a registry of 239 patients who wore 30-day loop recorders after discharge for cryptogenic ischemic stroke, researchers found that 24 % of all detected cases of occult atrial fibrillation were found in the final 10 days of 30-day monitoring (i.e., between days 20 and 30) [14]. Similarly, in a larger controlled trial of a similar patient population where 24-h Holter monitoring was compared to 30-day monitoring, 17 % of all cases of atrial fibrillation were detected in the final week of monitoring [8]. Although conventional wisdom suggests that longer monitoring may be useful for detecting rarer, potentially high-risk arrhythmias, such as ventricular arrhythmias and bradyarrhythmias, there is little empirical evidence on the impact and diagnostic yield of longer continuous monitoring for other clinically meaningful arrhythmias outside of atrial fibrillation. One advantage to the studied technology compared to typical 24- or 48-h Holter monitor systems is its longer continuous wear time up to 14 days, as well as its application without any long wires attached to distant electrodes. For longer monitoring periods, loop or event recorders have typically been the preferred technology, with the main disadvantage being that recordings are only stored if they meet pre-defined algorithms or for symptomatic triggers. A post-hoc investigation of the patient’s rhythm pre-or post-event cannot be done. Implantable loop recorders are occasionally used for very rare arrhythmia events, but these have the same limitations as loop and event monitors and also require a small surgical procedure to implant the device with its attendant risks.

For ventricular arrhythmias, we found that although the majority of arrhythmias are identified in the first 7 days, a significant proportion of arrhythmias were still detected in the 7 to 14-day monitoring window. This was more pronounced for non-sustained VT than sustained VT, with more than 20 % of non-sustained VT being identified in the 7–14 day window. Although we did not have detailed clinical characteristics for our patient population, in high-risk patients, such as those with cardiomyopathy, non-sustained VT can be a high risk marker that may warrant a change in treatment such as the consideration of an implantable cardioverter-defibrillator in certain patient populations (i.e., hypertrophic cardiomyopathy). Depending upon the clinical circumstances, both nonsustained and sustained VT often support the need for further diagnostic testing, such as the evaluation for structural heart disease or for cardiac ischemia. Similarly, although potentially high-risk bradyarrhythmias were less common, if they are not appropriately identified and treated, patients may suffer significant morbidity and excess mortality. The consideration of therapeutic interventions such as permanent pacemaker implantation is recommended by the joint American College of Cardiology, American Heart Association, and Heart Rhythm Society guidelines for the high-risk bradyarrhythmias evaluated in our study [15].

Our study had certain limitations. We did not have data on any changes in clinical management or patient outcomes following monitoring, so we were unable to delineate the direct clinical impact from the detection of arrhythmias found from the monitors in our study. We did not have information on all symptomatic triggers, and thus did not analyze the proportion of all symptomatic triggers that correlate to true arrhythmias. Patient information was limited to demographic characteristics, and data were unavailable on patients’ comorbidities, which could potentially help further risk stratify patients and allow for predictive modeling to help identify those most at-risk for high-risk arrhythmias. In addition, some bradyarrhythmias, such as asymptomatic sinus and AF pauses, may occur nocturnally in normal subjects. Further, we did not validate the data on the clinical indication for the ordered monitors, and differences among providers’ thresholds for ordering the monitors could have an impact on patient selection and arrhythmia yield. Finally, while average wear time was high, at least a quarter of patients wore the device for less than 7 days (25th percentile of 6.8 days), thus artificially reducing the yield of detected arrhythmias from 7 to 14 days. Thus, the actual yield of detected arrhythmias from days 7 to 14 or monitoring may be even higher than we observed.