Patients at Heightened Risk for Early Demise Following CRT
Patients at Heightened Risk for Early Demise Following CRT
While current guidelines advocate CRT in appropriate candidates with advanced heart failure before listing for advanced treatment options, some patients fail to respond to CRT potentially delaying more definitive therapies and possibly leading to inferior outcomes. Which patients will rapidly deteriorate despite CRT is unknown. In this study, we describe for the first time a simple clinical prediction rule to predict rapid deterioration following CRT in patients with advanced heart failure. The presence of at least 3 of 4 variables in the rule is associated with a 95% specificity for rapid deterioration following CRT.
While CRT represents one of the most important modern advances for the treatment of drug refractory heart failure, nonresponse in a relatively large number of patients continues to be problematic. Depending on one's definition of response, nonresponders to CRT may make up anywhere from 20% to 40% of patients who are currently implanted. The downgrade of non-LBBBs noted in the focused 2012 update, may cut the nonresponse rate somewhat but will almost certainly not eliminate it. While nonresponse to CRT is a major problem for all CRT recipients, it may pose even bigger risks to patients with more advanced disease who may have a limited window for therapeutic intervention. As such, clinicians who are treating patients with advancing, rapidly progressive heart failure must often decide between a trial of CRT versus a strategy of direct listing for LVAD and heart transplantation. There exists a paucity of data to help physicians make this decision.
While multiple predictive scoring systems have been developed to determine prognosis in ambulatory heart failure patients, none are adequate to determine whether to pursue or bypass CRT in patients with advanced heart failure. Most heart failure risk models in common use were derived before the era of routine CRT use. In addition, some indices were derived from a significant number of patients with an LVEF ≥40%. One exception, the HF-ACTION risk score, was derived from a large group of patients between 2003 and 2007; however, only 18% of patients in this cohort had a CRT device. No risk score for early demise has been derived exclusively from advanced heart failure patients receiving CRT.
All 4 variables in our rule have been shown in prior studies to be predictive of increased risk in patients with heart failure. Increased LVEDD is well known to predict an increased risk for poor outcomes in patients with advanced heart failure and is part of the Munich Score. Elevated creatinine levels have been shown to predict increased risk in patients with heart failure including, specifically, patients with a wide QRS undergoing CRT. The presence of a non-LBBB ECG pattern may portend an increased risk for poor outcomes following CRT due to a lower chance of response from CRT, as well as being a marker of more advanced left ventricular dysfunction. Finally, absence of β-blocker usage in this population followed almost entirely be dedicated heart failure specialists likely represents intolerance to the medication rather than omission. Therefore, this variable likely represents more advanced pump dysfunction. Interestingly, age was not a factor in predicting early demise. In subgroup analysis from the MADIT-CRT trial, elderly patients derived a significant clinical benefit; therefore, it is not surprising that age was not a risk factor for poor outcomes in our study.
The presence of at least 3 variables in our model had a sensitivity of 21% and a specificity of 95% for early demise. While a rule providing both sensitivity and specificity would have been ideal, specificity in this instance is significantly more important so as not to withhold CRT from patients who may benefit. Still, we feel the specificity of 95% with 3 or more variables is sufficiently high to provide clinicians a valuable tool in deciding to pursue or bypass CRT. It is important to note that many patients with a single or even 2 risk factors may have quite good outcomes following CRT. Therefore, the decision to bypass CRT should be made in the context of patients with multiple risk factors (3 or all 4). In patients who are not candidates for advanced heart failure therapies and have a high risk for early demise, the decision to implant a CRT device is best left to physician judgment and determined on an individual basis.
Our study has several limitations. The retrospective nature cannot account for all confounders despite our best efforts to identify important baseline differences. Patients in our cohort come from a single tertiary-care center and, therefore, may not be representative of patients presenting to other centers. The population of patients undergoing CRT with early demise is relatively small but represents the largest such cohort in the literature. The prediction rule was not validated in a different cohort. While bootstrapping methods were used to improve external validity, prospective validation is needed to determine the rule's true utility. Not all factors described in the medical literature associated with response were included. For example, measurements of right ventricular function were not included. Given the retrospective nature of this study, objective quantifications of right ventricular function were not available and hence were left out of the analysis.
In conclusion, severe left ventricular dilatation, the presence of a non-LBBB morphology, renal dysfunction, and a lack of or intolerance to β-blockers are associated with early demise following CRT. In patients with at least 3 of these factors, bypassing CRT with an early adoption of advanced heart therapies may be considered.
Discussion
While current guidelines advocate CRT in appropriate candidates with advanced heart failure before listing for advanced treatment options, some patients fail to respond to CRT potentially delaying more definitive therapies and possibly leading to inferior outcomes. Which patients will rapidly deteriorate despite CRT is unknown. In this study, we describe for the first time a simple clinical prediction rule to predict rapid deterioration following CRT in patients with advanced heart failure. The presence of at least 3 of 4 variables in the rule is associated with a 95% specificity for rapid deterioration following CRT.
While CRT represents one of the most important modern advances for the treatment of drug refractory heart failure, nonresponse in a relatively large number of patients continues to be problematic. Depending on one's definition of response, nonresponders to CRT may make up anywhere from 20% to 40% of patients who are currently implanted. The downgrade of non-LBBBs noted in the focused 2012 update, may cut the nonresponse rate somewhat but will almost certainly not eliminate it. While nonresponse to CRT is a major problem for all CRT recipients, it may pose even bigger risks to patients with more advanced disease who may have a limited window for therapeutic intervention. As such, clinicians who are treating patients with advancing, rapidly progressive heart failure must often decide between a trial of CRT versus a strategy of direct listing for LVAD and heart transplantation. There exists a paucity of data to help physicians make this decision.
While multiple predictive scoring systems have been developed to determine prognosis in ambulatory heart failure patients, none are adequate to determine whether to pursue or bypass CRT in patients with advanced heart failure. Most heart failure risk models in common use were derived before the era of routine CRT use. In addition, some indices were derived from a significant number of patients with an LVEF ≥40%. One exception, the HF-ACTION risk score, was derived from a large group of patients between 2003 and 2007; however, only 18% of patients in this cohort had a CRT device. No risk score for early demise has been derived exclusively from advanced heart failure patients receiving CRT.
All 4 variables in our rule have been shown in prior studies to be predictive of increased risk in patients with heart failure. Increased LVEDD is well known to predict an increased risk for poor outcomes in patients with advanced heart failure and is part of the Munich Score. Elevated creatinine levels have been shown to predict increased risk in patients with heart failure including, specifically, patients with a wide QRS undergoing CRT. The presence of a non-LBBB ECG pattern may portend an increased risk for poor outcomes following CRT due to a lower chance of response from CRT, as well as being a marker of more advanced left ventricular dysfunction. Finally, absence of β-blocker usage in this population followed almost entirely be dedicated heart failure specialists likely represents intolerance to the medication rather than omission. Therefore, this variable likely represents more advanced pump dysfunction. Interestingly, age was not a factor in predicting early demise. In subgroup analysis from the MADIT-CRT trial, elderly patients derived a significant clinical benefit; therefore, it is not surprising that age was not a risk factor for poor outcomes in our study.
The presence of at least 3 variables in our model had a sensitivity of 21% and a specificity of 95% for early demise. While a rule providing both sensitivity and specificity would have been ideal, specificity in this instance is significantly more important so as not to withhold CRT from patients who may benefit. Still, we feel the specificity of 95% with 3 or more variables is sufficiently high to provide clinicians a valuable tool in deciding to pursue or bypass CRT. It is important to note that many patients with a single or even 2 risk factors may have quite good outcomes following CRT. Therefore, the decision to bypass CRT should be made in the context of patients with multiple risk factors (3 or all 4). In patients who are not candidates for advanced heart failure therapies and have a high risk for early demise, the decision to implant a CRT device is best left to physician judgment and determined on an individual basis.
Our study has several limitations. The retrospective nature cannot account for all confounders despite our best efforts to identify important baseline differences. Patients in our cohort come from a single tertiary-care center and, therefore, may not be representative of patients presenting to other centers. The population of patients undergoing CRT with early demise is relatively small but represents the largest such cohort in the literature. The prediction rule was not validated in a different cohort. While bootstrapping methods were used to improve external validity, prospective validation is needed to determine the rule's true utility. Not all factors described in the medical literature associated with response were included. For example, measurements of right ventricular function were not included. Given the retrospective nature of this study, objective quantifications of right ventricular function were not available and hence were left out of the analysis.
In conclusion, severe left ventricular dilatation, the presence of a non-LBBB morphology, renal dysfunction, and a lack of or intolerance to β-blockers are associated with early demise following CRT. In patients with at least 3 of these factors, bypassing CRT with an early adoption of advanced heart therapies may be considered.
Source...