Device implantation and complications: time to recalibrate our expectations?

There has been a worldwide increase in the number of patients receiving implantable pacemakers, defibrillators, and resynchronization therapy devices. Although, there is evidence from previous studies and clinical trials that implant complication rates modestly increase with the complexity of devices, these reports may not be entirely reflective of the real-world experience. This is of particular importance when discussing therapeutic options with our patients and keeping the goal of providing cost-effective ‘high value’ care.The relevance of institutional and country-specific metrics is further augmented by the rapidly evolving healthcare landscape and the need to administer personalized care via accurate assessments of the risk– benefit of therapeutic strategies such as device implantation. Device therapy is resource intensive, and its cost-effective use should be guided by rea-world data on outcomes, risks, and complications. Kirkfeldt et al. now present the first nationwide, populationbased study comparing pacemakers and high voltage device implantation procedures. The authors describe complication rates for implantable cardiac devices in 5918 consecutive patients who underwent device implantation in Denmark from May 2010 to April 2011. The patients were enrolled prospectively with the aim of capturing pre-defined data. Importantly, although the data capture in this study was retrospective, rigorous ascertainment of complications was accomplished by chart review of every single subject. The results are provocative, showing a higher than expected rate of any complication (9.5%). Several known risk factors for device complications were confirmed, including higher rates for implantable cardioverter-defibrillators (ICDs) compared with pacemakers, and higher risk associated with more complex procedures such as dual-chamber and biventricular devices (Figure 1). The higher infectious risk associated with generator replacement over primary implants was observed. Other risk factors included female sex, low body mass index, and operators with low annual procedure volumes. The strengths of this report include thorough evaluation of all complications and inclusion of all patients undergoing cardiac device implants during the study period. However, there are some limitations inherent in the study design. Primary among them is the lack of knowledge of patient-specific factors that are known to influence risk of procedural complications (e.g. renal function, diabetes, etc). Furthermore, understanding the indications for the procedures is an important part of the patient profiles, as these, in turn, can influence outcomes and complications. Without controlling for such patient-specific factors, it is difficult to know, for instance, whether low volume operators had intrinsically higher complication rates, or merely suffered from an unfavourable case mix. Notably, the results show that in a real-world setting, complication rates are higher than seen in the large clinical trials. How can we account for this difference and should we be concerned? There are several reasons why complications might increase in the real world compared with clinical trials. One reason is because clinical trials are usually performed athigh volume centreswith more experienced operators, while the real-world experience may include a number of low volume operators and centres. However, differences in operator and centre experience may not entirely account for the observed differences. In a similar study to this, Al-khatib et al. found that subjects matched by propensity score from the US National Cardiac Device Registry (NCDR) had similar outcomes to subjects from MADIT-II and SCD-HeFT, suggesting that when patient-specific risk factors are controlled, real-world operators do as well as those involved in clinical trials for ICD implant outcomes. The data capture in this study was clearly exemplary, with no patient lost to follow-up. Could it be that complication rates were simply better captured by the rigorous chart review performed in this study? While this is a possibility, the prospective pivotal clinical trials performed for cardiac devices are highly rigorous with superb follow-up and are unlikely to have missed a large percentage of procedural complications. We are left then with the patients themselves. Could it be that in the real-