Novel variables aid prediction of high post-PCI mortality risk
MedWire News: Patients at very high risk for death after undergoing percutaneous coronary intervention (PCI) may be identified by assessing preprocedural patient variables that are not currently included in existing coronary intervention risk models, US researchers report.
"The new covariates were termed 'compassionate use' (CU) criteria and... included coma on presentation for emergent PCI, use of a percutaneous ventricular assist device or extracorporeal bypass, and cardiopulmonary resuscitation at the initiation of the procedure," state Frederic Resnic (Brigham and Women's Hospital, Boston, Massachusetts) and colleagues.
They explain that improving the ability of risk models to identify extremely high-risk patients is likely to increase the confidence with which clinicians can gauge the probable outcomes of patients referred for PCI.
The study, published in the Journal of the American College of Cardiology, involved 5588 patients scheduled for PCI following admission to a Massachusetts hospital for cardiogenic shock or ST-elevation myocardial infarction.
Of these patients, 96 (1.7%) had at least one CU criterion at the start of PCI.
Resnic and team found that the presence of one or more CU criteria produced a 27.3-fold increase in risk for death during hospital admission compared with the absence of all CU criteria.
Furthermore, the discriminatory ability of in-hospital mortality prediction models (developed using regression models and assessed with areas under the receiver-operating characteristic curves) improved from 0.87 to 0.90 prior to and after the inclusion of CU criteria, respectively.
When all patients were assigned to one of five risk strata, according to their risk-model score, Resnic and colleagues noted that the use of CU criteria improved overall stratification accuracy by 8.7% compared with stratification without the inclusion of CU variables.
In a related commentary, Eric Peterson, from Duke Clinical Research Institute in Durham, North Carolina, USA, wrote: "Resnic et al have identified a rare but important set of high-risk factors not previously collected in national registries."
He expressed concern that the identification of high-risk patients may deter interventionalists from attempting high-risk PCI cases, and therefore, called for ongoing evaluation of the influence of profiling methods on case selection by healthcare providers.
MedWire (www.medwire-news.md) is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2011
By Lauretta Ihonor