Abstract Background Proactive patient monitoring is of paramount importance in effective management of heart failure (HF) patients. Cardiac implantable electronic devices (CIEDs) used in HF patients are able to derive long-term trends in physiologic parameters and provide timely warning to clinicians. Little is known, however, on the real-world experience with device-generated HF risk-stratifying algorithms. Purpose Heart Failure Risk Score (HFRS) takes into account nine parameters and is calculated automatically based on long-term clinical trends. Remote transmissions provide information on the risk of HF event in next 30 days categorized as low, medium or high based on a maximum daily risk status in prior 30 days. We aimed to evaluate the ability of HFRS to alert HF specialists on the actual HF risk status. Methods The prospective registry included all patients with CIEDs featuring integrated Heart Failure Risk Score (HFRS) followed via Medtronic CareLink remote monitoring system and enabled for Co-management (CM) from May 2015 to August 2019 in a tertiary centre. High HFRS does not trigger automatic alert transmission. Study follow-up spanned between start of CM and last transmission in 2019. Inclusion criteria were CRT-D in situ, active Home Monitor, switched on OptiVol 2.0 remote alert and transmission data available on CareLink following study period completion. Transmissions were scheduled 3-monthly. Results Out of 229 consecutive patients, 132 met study criteria. Mean age was 74±10 years, 18% were female. Median follow-up duration was 2.7 years (IQR 1.3). Total number of transmissions was 2652, median per patient was 18 (IQR 13); scheduled, unscheduled and care alerts constituted 42%, 44% and 14%, respectively. One third of transmissions were automatically sent for CM review. There were 398 high HFRS episodes. OptiVol fluid index was below the threshold throughout 128 (32%) episodes. Missed episodes (not transmitted within 30 days from the final day of high HFRS) amounted to 130 (33%) and the reasons behind this included OptiVol alerting before the first day of high HFRS or persistently elevated when HFRS changed from low/medium to high (52%), low OptiVol index during the episode (38%) or other (10%). Median duration of high HFRS was 7 days (IQR 12, range, 1–187). Among timely picked-up high HFRS episodes, 38% were transmitted during the relevant episode and 62% afterwards with median delay of 10 days (IQR 15) from the last day of high HFRS; 21% of transmissions showing high HFRS were not highlighted for CM review which correlated with low OptiVol index, P<0.001. The factors contributing to high HFRS included: raised OptiVol (60%), patient activity (83%), AT/AF (46%), ventricular rate (VR) during AF (6%), % of VP (40%), shocks (2%), treated VT/VF (2%), night VR (72%) and HR variability (34%). Conclusions In a real-world clinical setting high HFRS was frequently under-reported. The investigation into clinical implications is warranted. Funding Acknowledgement Type of funding source: Private grant(s) and/or Sponsorship. Main funding source(s): Our department has benefited from unrestricted grants from Boston Scientific and Medtronic Inc during the last 5 years.