A predictive scoring system that combines both echocardiography and clinical data can reliably estimate three-year death rates in critical care unit (CCU) patients, according to research from the Mayo Clinic in Rochester, MN. The model is built on the fact that more heart patients are dying from noncardiovascular diseases.
The scoring model was comparable to or better than other patient mortality assessment methods in these patients -- and it's also simpler and easier to use, according to medical student Prakriti Gaba and colleagues.
"The clinical implications of our score may be drastic, not only in terms of having a more tangible way in which CCU clinicians will be able to assess mortality, but also in terms of cutting down unnecessary healthcare costs due to better stratification of CCU resources," Gaba told AuntMinnie.com.
The researchers shared their findings in a poster at the recent American Society of Echocardiography (ASE) annual meeting.
Limited resources in the CCU
Physicians are increasingly encountering complex and diverse patient populations in the CCU, and it could be helpful to have a predictive scoring system to identify and stratify patients who would benefit the most from the limited medical resources available, Gaba said.
Although patient mortality assessments such as the Acute Physiology and Chronic Health Evaluation (APACHE) III score, the Sequential Organ Failure Assessment (SOFA) score, and the Simplified Acute Physiology Score (SAPS) II are available, they were not developed in a CCU population and, consequently, have limited prognostic ability in these patients.
"Moreover, routine echocardiography has become increasingly common," she said. "This coupled with the fact that echo parameters such as [left ventricular ejection fraction] and [right ventricular systolic pressure] have been shown to be prognostic of several cardiovascular illnesses encouraged our incorporation of these factors (coupled with relevant clinical parameters) into a prognostic score to assess CCU mortality."
The Mayo group developed its model from 10,411 consecutive patients who had been admitted to the CCU at the Mayo Clinic between January 2004 and April 2011. After excluding 869 patients who declined authorization of their records and 1,323 other patients who did not have available echocardiographic data, the researchers were left with an initial cohort of 8,219 patients. This group had a complete set of echocardiographic parameters consisting of left ventricular ejection fraction, left ventricular diastolic function, right ventricular function and size, right ventricular systolic pressure, and stroke volume index.
Left ventricular ejection fraction and right ventricular systolic pressure -- the only factors found to be independently associated with three-year mortality -- were ultimately included in the model. The researchers then included the most common relevant clinical parameters from the Charlson Comorbidity Index (CCI): age, diabetes mellitus, chronic pulmonary disease, and chronic kidney disease.
After testing the final combined scoring model on a separate cohort of 2,424 patients, the researchers compared their model's prognostic ability with the APACHE III, SOFA, and SAPS II scores for the same patients using receiver operator characteristics (ROC) analysis.
| Mayo CCU scoring model vs. other models | ||
| Model | Area under the ROC curve | 95% confidence interval |
| Mayo scoring | 0.76 | 0.74-0.78 |
| APACHE III | 0.76 | 0.74-0.78 |
| SOFA | 0.72 | 0.70-0.74 |
| SAPS II | 0.72 | 0.70-0.74 |
| Charlson Comorbidity Index | 0.71 | 0.69-0.73 |
A better, simpler scoring system
Mayo's proposed CCU scoring model performed comparably to APACHE III (p = 0.929) and offered a statistically significant improvement in predictive ability over SOFA (p = 0.014), SAPS II (p = 0.010), and CCI (p = 0.00001), according to the researchers. What's more, it doesn't require as many variables as the other scoring methods.
"Our score is simple to use with only six parameters -- a significant reduction in inputs relative to existing scores -- and thus can easily be applied to clinical practice," Gaba said. "The aim is to have CCU clinicians use the score when admitting patients to the CCU in order to assess patient prognosis at the start of admission, allowing for accurate direction of care."
Gaba noted that noncardiovascular diseases are becoming increasingly common as causes of death in patients with myocardial infarction and heart failure.
"As a result, incorporating significantly relevant comorbidities that are independently associated with long-term mortality into a score seemed like a logical way in which to develop a meaningful prognostic score," she said.
While the model is experimental at this point, the researchers hope to verify it in a few other tertiary care centers to increase sample size and improve its prognostic ability, "after which we hope to have CCU clinicians incorporate it into their daily patient regimen," she said.
