The IMPACT – IMproving Paediatric and Adult Congenital Treatment – Registry is a catheterisation registry of all children and adults with CHD undergoing diagnostic and interventional cardiac catheterisation. Following discussions regarding the potential of this registry in 2006, a pilot project was launched in 2009 and the registry was implemented in December, 2010. As there were very few sites actually entering data in 2010, what follows is a report of data collected for nine quarters beginning in January, 2011 and ending in March, 2013.
Registry data are useful in identifying, amongst other things, benchmark metrics for success rates and complications for specific procedures. The complexity and diversity of congenital cardiac diagnoses makes it difficult to collect large numbers of similar procedures from any single centre. Single-centre data that are published often are skewed towards those centres with the best outcomes and/or larger centres. Although this may set a standard, it does not necessarily reflect care that is provided in the general community. The intent of the IMPACT Registry is to gather information from as many centres as possible performing diagnostic and interventional catheterisation procedures on all children and adults with CHD. It is hoped that through the registry we will benchmark performance, identify best practices, areas requiring improvement, and ultimately improve patient outcomes. As with other National Cardiovascular Data Registries, participation is voluntary and institutions need not submit data.
Previous congenital cardiac databases have concentrated mainly on interventional procedures.Reference Stanger, Cassidy, Girod, Kan, Lababidi and Shapiro 1 – Reference Forbes, Garekar and Amin 8 In its design, the IMPACT Registry aims to collect catheterisation data on all children and adult congenital cases, including all diagnostic, biopsy, and interventional procedures. More in-depth data were obtained on the following six specific interventions: device closure of atrial septal defect; device closure of patent ductus arteriosus; pulmonary valvuloplasty for pulmonary stenosis; aortic valvuloplasty for aortic stenosis; interventions – angioplasty and stenting – on coarctation of the aorta; and central pulmonary artery stenting. An important difference between this registry and others is the inclusion of all catheterisation cases with unrestricted participation. During the development of the registry, the steering committee felt that diagnostic catheterisations should be separated from interventional ones, as they would be less complex. To test this hypothesis, data are presented for diagnostic-only procedures for comparison with interventional ones.
Methods
The IMPACT Registry® is an initiative of the American College of Cardiology Foundation with partnering support from The Society for Cardiovascular Angiography and Interventions and the American Academy of Pediatrics, and has been previously described by Martin et al.Reference Martin, Beekman and Ing 9 The Registry collects data for use in the development of performance and quality metrics, quality improvement programmes, and peer-reviewed outcome research focussed on children and adults with CHD, who are undergoing diagnostic catheterisations and catheter-based interventions. Demographic, clinical, procedural, and institutional data elements were collected at participating centres and entered via either a web-based platform or software provided by American College of Cardiology-certified vendors, and were collected in a secure, centralised database. The IMPACT Registry has in place a data quality programme consistent with that described for the IMPACT Registry® Reference Messenger, Ho and Slattery 10 . A comprehensive description of IMPACT Registry data elements and definitions is available at http://www.ncdr.com. The IMPACT Registry research studies are determined to meet the definition of research not requiring informed consent, as patient information is collected anonymously and without unique patient identifiers.
Data collection forms submitted to the National Cardiovascular Data Registry were analysed for completeness, and those with adequate information were used for the purposes of statistical analysis in the registry. The data are presented by quarter of event to identify the growth of the registry in terms of number of centres enrolled and number of events in a quarter and on a cumulative basis. This report focusses on general data for patients undergoing diagnostic cardiac catheterisation as well as interventional catheterisation procedures. Patients having one of the six specific interventions and those with multiple interventions or interventions other than the six specific ones were addressed in different reports.
Patients undergoing catheterisation procedures were divided into four age groups as follows:
Newborns: <30 days; Infants: ⩾30 days, ⩽1 year; Children: >1 year, ⩽18 years; Adults: >18 years.
For each age group, data for patients undergoing diagnostic and interventional catheterisations were analysed for the following major adverse events: need for a permanent or temporary pacemaker; cardiac arrest; tamponade requiring pericardial drainage); embolic stroke (within 72 hours); and unplanned cardiac surgery due to catheterisation complications.
Other adverse events also reported and analysed include the following: arrhythmias, including any arrhythmia as well as those requiring cardioversion or antiarrhythmic medication; new heart valve regurgitation; air embolus; device malposition or thrombus requiring surgery; device embolisation requiring retrieval by catheterisation or surgery; new requirement for dialysis; airway event requiring intubation; event requiring extracorporal membrane oxygenation; event requiring left ventricular assist device; bleeding event including bleeding event at access site, haematoma at access site, retroperitoneal bleeding, gastrointestinal bleeding, genital-urinary bleeding, and other bleeding; red blood cell or whole blood transfusion; other events; unplanned vascular surgery; and unplanned other surgeries and subsequent cardiac catheterisations.
Adverse events are reported if they occur within 30 days after the procedure, except for embolic stroke, bleeding events, and transfusions, which are reported only within 72 hours following the procedure. Bleeding events are characterised as follows: haemoglobin drop of ⩾3 g/dl; transfusion of whole blood or packed red blood cells; procedural intervention/surgery at the bleeding site to reverse/stop or correct the bleeding, such as surgical closures/exploration of the arteriotomy site, balloon angioplasty to seal an arterial tear, and endoscopy with cautery of a gastro-intestinal bleed. If a subsequent catheterisation occurs before the time limit for collecting adverse events, future events are attributed to the subsequent catheterisation.
Adverse events are not necessarily caused by the catheterisation procedure itself as other events, including surgery, may have occurred between catheterisation and the adverse event. Adverse events are reported as a percentage of patients in each age group who have an event during or after a catheterisation procedure, whereas the number of deaths is reported/episode of care. Deaths are dealt with separately, and like the other adverse events the data indicate only that mortality occurred during the admission and not that it is necessarily attributable to the catheterisation procedure itself.
Descriptive statistics for catheterisation procedures are presented. All the categorical variables are reported as frequency and/or percentage and they are compared among age groups using the χ2 test or Fisher’s exact test. All the results including the group comparison and the odds ratios are based on univariate analyses. All statistical analyses were conducted using SAS Version 9.3 (SAS Institute, Cary, North Carolina, United States of America).
Results
Registry participation has grown from 12 sites (Q1 2011) to 73 sites (Q1 2013). Currently there are 91 sites in the IMPACT Registry. Along with the growing number of sites, the total number of event submissions has increased from 848 (Q1 2011) to 19,797 (Q1 2013) (Fig 1).
Growth of the IMPACT Registry by number of participating centres per quarter and cumulative number of cases over time (times 100).
Of the 19,797 procedures analysed in the registry, 7010 (35.4%) procedures were for diagnostic “only” catheterisations; 23.8% of the procedures were for one of the specific six isolated interventions being evaluated; and 40.7% of the procedures were for “other” reasons. These include endomyocardial biopsies, interventions other than the six cited – for example, pulmonary artery angioplasty – and multiple procedures – for example, atrial septal defect device closure and patent ductus arteriosus occlusion during the same procedure, or a pulmonary artery stent as well as peripheral pulmonary artery dilations (Table 1).
All procedures are categorised by age group and procedure type.
Procedures categorised by age group and procedure type. Percentages refer to the percentage of the specific procedure type compared with the total number of procedures or total for that age group
Tables 2a and b summarise the results of data collected on procedural status, sedation, and adverse events for diagnostic-only and interventional procedures. Table 2c compares the same data elements of the two groups. There were some differences in procedural status and sedation with interventional cases being more elective and more likely to use anaesthesia support.
Procedural characteristics and adverse events for patients having a diagnostic catheterisation.
ECMO=extracorporeal membrane oxygenation; IM=intramuscular; IV=intravenous; LVAD=left ventricular assist devices; RBC=red blood cells
Arrhythmias and bleeding events are subcategorised. Results are shown as the total for the cohort and broken down into age categories Percentages refer to the percent within the age group
Procedural characteristics and adverse events for patients having an interventional catheterisation.
ECMO=extracorporeal membrane oxygenation; IM=intramuscular; IV=intravenous; LVAD=left ventricular assist devices; RBC=red blood cells
Arrhythmias and bleeding events are subcategorised. Results are shown as the total for the cohort and broken down into age categories. Percentages refer to the percent within the age group
Procedural characteristics and adverse events for patients having a diagnostic catheterisation compared with an interventional one.
ECMO=extracorporeal membrane oxygenation; IM=intramuscular; IV=intravenous; LVAD=left ventricular assist devices; RBC=red blood cells
Arrhythmias and bleeding events are subcategorised
Major adverse events not including death occurred in 1.4% of diagnostic catheterisation procedures, (major adverse event/catheterisation lab visit), and in 1.2% of interventional procedures. Major adverse events were most common in neonates (4.7% of diagnostic catheterisation; 4.2% of interventional procedures) with decreasing prevalence in infants (2.1% diagnostic versus 1.9% of interventional procedures). Adult patients had a significantly higher prevalence of major adverse events (1.4% of diagnostic versus 1.1% of interventional procedures) than children and adolescents (0.7:0.8%,) (Tables 2a and b). These differences in major adverse events among age groups were statistically significant (p value<0.001). Any adverse event occured in 10.0% of diagnostic procedures and 11.1% of interventional cases (p=0.013), also being particularly prevalent in newborns and infants (Tables 2a–c).
Vascular access: venous access sites other than femoral were reported in 12.8% of diagnostic cases and 22% of interventional ones. Non-femoral arterial access was used in 2.3% of diagnostic cases – 0.34% using the umbilical artery – and 1.8% of interventional procedures – 0.38% using the umbilical artery. Analysing the use of non-femoral or umbilical arterial access, the carotid artery is most commonly used in newborns, infants, and children (97 procedures carotid, 31 radial), while the radial artery is used most commonly in adults (7 carotid, 62 radial).
All cause hospital mortality for patients having a catheterisation during their admission was 2.1%; 12.0% for newborns, 4.6% infants, 0.7% children, and 0.8% adults based on available data in 18,956 hospital admissions with the discharge status known (Table 3). Of the 395 deaths reported, 16 occurred in the catheterisation laboratory at the time of the procedure; 20.3% of patients who ultimately died, died within 3 days following the; procedure; 63.5% died after 7 days, with 43% of these patients dying after 2 weeks or more following the last catheterisation procedure (Table 4). The primary cause of death is listed in Table 4. Cardiac causes account for 58.7% of deaths with pulmonary causes another 17.7% (Table 5). The IMPACT Registry currently does not account for other procedures – for example, cardiac and non-cardiac surgical procedures – which might have occurred either before or after a catheterisation procedure, making attribution of death or AE to the catheterisation difficult.
Categorisation of death during the episode of care.
Time from catheterisation to death.
Primary cause of death.
Of the potential risk factors and genetic syndromes that might have contributed to adverse events and poor outcomes, only diabetes mellitus, renal insufficiency, seizure disorder, and single ventricle were statistically associated with risk of a major adverse event. Genetic syndromes were not associated with any increased risk of a major adverse event.
Discussion
This harvest of over 2 years of data collection demonstrates some of the trends we previously suspected. Diagnostic catheterisations now account for approximately only one-third of all procedures. This is similar to data reported by others.Reference Bergersen, Gauvreau and Marshall 6 Previous reports have focussed on procedural characteristics and adverse events of interventional procedures. As IMPACT collects data on all procedures, we are able to separate diagnostic procedures from the interventional ones. What we are seeing, however, is little difference in characteristics or adverse events between the two groups, although there are significant differences in adverse events between age groups.
Anaesthesiology support is being used in the majority of cases in most institutions, 84.1% for diagnostic procedures and 87.8% for interventional ones. This may be due to a combination of operator preference as well as hospital and national sedation policies, but regardless of why the support is used, the rate of airway events requiring intubation is <1%.
Although vascular access is still predominately from the femoral approach, non-femoral sites for venous access are used in 12.8% of diagnostic and 22% of interventional cases. Non-femoral arterial access is low with the carotid artery often being preferred in children compared with the radial artery in adults. As vascular access issues are not necessarily known before undertaking the catheterisation procedure, this may question whether physicians performing these procedures should be versed in more than just femoral access techniques.
Major adverse events are relatively uncommon (1.4% of diagnostic, 1.2% of interventions), but are highest in the newborn age group. Adverse events occurred in 10.0% of diagnostic cases and 11.1% of interventional cases. Reports of adverse event rates vary widely and range from 20%Reference Bergersen, Marshall and Gauvreau 11 for any adverse event to 1.9–7.4% for different institutions when comparing high-severity adverse events Reference Bergersen, Gavreau and Foerster 7 . Given the different definitions and methodologies used, it is not possible to compare results, but it would appear that the results of this multi-institutional registry are within range of other reports.
The IMPACT® Registry in its current form is unable to attribute adverse events to the catheterisation procedure or other factors. This is particularly true for death before discharge. The mortality discussed is all cause mortality during the episode of care and is not necessarily attributable to the catheterisation procedure. Death is accounted for only as an event in the catheterisation laboratory or as the discharge status of the patient, regardless of other events, such as surgery, occurring during the hospital stay. Table 5 shows that nearly 80% of reported deaths occurred >72 hours after the catheterisation procedure. The fact that mortality risk is associated with patient age but not procedural factors (intervention versus diagnostic) strongly suggests that much of mortality risk is due to patient factors rather than the catheterisation procedure. Nonetheless, we believe that these data do portray a realistic picture of the overall mortality risk faced by patients requiring a cardiac catheterisation.
This first harvest of data has revealed several issues pertinent to the next version of the registry. As there are few differences in procedural characteristics and adverse events in diagnostic compared with interventional procedures, it is clear that a model of procedural complexity and risk stratification will be necessary to understand procedural risk and compare results. The current list of adverse events, major adverse events, and co-morbidities need revision. They were initially adopted from the National Cardiovascular Data Registry for Percutaneous Catheter Interventions, and although they provided a starting point they failed to capture some of the differences and definitions of paediatric problems compared with those seen in adults. The intimate relationship of congenital cardiac catheterisation and surgery, with surgery often occurring shortly after a catheterisation procedure, will be addressed in the next version of IMPACT. Currently, adverse events are captured for 72 hours to 30 days, depending on the event, following the catheterisation procedure or terminated with the next catheterisation procedure or discharge. With the revision of the registry, the next surgical procedure will be included as a stopping point for capturing catheterisation-related events, unless the surgical procedure is a result of a catheter complication. The intent of the registry as constructed was to capture adverse events that impact patient care; however, the consequence of this approach is the difficulty in attributing those events to the catheterisation procedure. Attribution can be very subjective and may vary widely from centre to centre, particularly for patients who have long hospital stays unrelated to a catheterisation procedure. We hope to institute a shorter window of time for collecting adverse events post-procedure, as this may be seen by the operators as more likely being related, or possibly related, to the catheterisation procedure, thus driving quality improvement while at the same time allowing for more consistent input of data from all centres.
Finally, although data collection is currently limited to hospital discharge, we hope to extend this to long-term follow-up for some specific procedures with future versions.
The intent of the IMPACT Registry is to gather data to set national benchmarks for diagnostic and certain specific interventional procedures. Future versions of IMPACT will address risk stratification, attribution of adverse events, and long-term follow-up of some interventions. Our hope is that the information generated will allow creation of recommendations to guide improvements in catheter intervention. Although adverse events as currently reported are not completely attributable to the catheterisation procedure itself, it is clear that newborns and infants who require a catheterisation procedure are at much higher risk for adverse events, including death, during their hospital admission than are older patients.
Although the IMPACT Registry was conceived and implemented to improve catheterisation outcomes, it should be stated unequivocally that, as currently constructed, centre-specific IMPACT data are not yet ready for comparisons between centres. Until risk adjustment and case complexity are available along with validated auditing and adjudication procedures, any attempt to compare variances between centres should be considered premature. First and foremost, the data reflect global events during an episode of care (hospital stay), not just what happens during, or as a consequence of, the index procedure. The data do, however, allow centres to look at their specific reports, and in a broad sense compare themselves to aggregate data. What may appear as a higher rate of major adverse events or death may be a case selection rather than a quality issue, some referral centres taking on cases others would not. These issues are not unique to IMPACT but are a growing concern amongst all registries, especially those requiring public reporting. It is, therefore, paramount that future versions of IMPACT address these issues.
Limitations of this report reflect current limitations of the IMPACT Registry. The data presented here were self-reported by centres and not audited, although there is currently a pilot audit project being performed. As mentioned above, there is no attribution of adverse events to the procedure, case complexity, or risk adjustment, making inter-institutional comparisons difficult.
Acknowledgements
The authors thank the staff of the American college of Cardiology, National Cardiovascular Data Registry, and Saint Luke’s Mid Atlantic Heart Institute for their assistance with the analysis and preparation of this manuscript.
Financial Support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors. The National Cardiovascular Data Registry contracts with Saint Luke’s Mid Atlantic Heart Institute for data analysis.
Conflicts of Interest
None.
