November 9, 2010
Table 4. ACC/AHA Stages of Heart Failure
A: At high risk for HF but without structural heart disease or symptoms of HF
B: Structural heart disease but without signs or symptoms of HF C: Structural heart disease with prior or current symptoms of HF D: Refractory HF requiring specialized interventions
Novel Risk Stratification: Low-Risk, Not High-Risk Markers Are Necessary Previous and ongoing research continues to identify individual markers of high risk associated with adverse events. The recurrent theme is obvious—hypotension, hyponatremia, renal dysfunction, increased troponin levels, and elevated natriuretic peptides all portend a poor prognosis.12,31,34,103,104,107,126,142,212,213 Unfortunately, markers of high risk rarely impact acute decision making, especially when prognosticating for events 6 to 12 months in the future. Although we know that these markers identify patients at risk for subsequent events, how does this impact disposition decisions? When risk is not immediate (ie, in-hospital morbidity or mortality), such markers have little bearing on the administration of acute therapy or the triage level for inpatient care. Stated another way, when emergency physi- cians are already admitting 4 of every 5 patients with AHFS to the hospital, will markers of high risk really alter practice patterns? Such data may prompt initiation of life-saving thera- pies such as -blockers or ACE inhibitors before hospital discharge, but these efforts would only modify intermediate- to long-term risk.
In essence, the absence of high-risk markers does not, by default, define a low-risk patient. Decision making for this large cohort of patients without high-risk features (ie, those with normal troponins, serum sodium, and renal function) has not been well studied. Can they be safely discharged directly from the ED or should they be managed in an observation unit? What if they have poor social support or lack access to timely outpatient follow-up? Biomarkers have emerged over the past decade as an effective means of stratifying patients with AHFS and, to varying degrees, may be useful for determi- nation of immediate or short-term risk. According to Morrow and de Lemos for a biomarker to be clinically useful it must meet the following 3 criteria: (1) accuracy on repeated measurements and available at a reasonable cost, (2) provision of additional information not already available from careful clinical assess- ment, and (3) the measured level should aid in decision mak- ing.214 Millions of dollars are spent and many papers are published in an attempt to delineate criteria 1 and 2; however, when 80% of patients are ultimately admitted, it appears that few if any AHFS prognostic biomarkers have fulfilled criteria 3 in terms of risk stratification. We clearly need to identify sensitive, meaningful markers with strong negative likelihood ratios that can identify patients who are truly at low risk for adverse events and can be safely discharged home.
Predictive Instruments May Be the Answer Although physicians and nurses exhibit intermediate accuracy for prediction of postdischarge death, their ability to estimate other metrics of risk such as need for subsequent rehospitaliza- tion is poor.215 Given the heterogeneity of the AHFS population,
Downloaded from http://circ.ahajournals
it is unlikely that any single biomarker will supersede others to such a degree that it will be the sole discriminator of discharge eligibility. Predictive instruments represent the most likely method of successfully defining low-risk AHFS patients. “Med- ical decision making” is the science of statistically examining detailed clinical data to develop mathematical models or predic- tive instruments to guide appropriate clinical care of patients with complex diseases.216–219 By accounting for commonly overlooked factors such as socioeconomic status and healthcare access, such predictive instruments can reduce the margin of error and increase the likelihood that clinicians will successfully identify those who are truly at low risk. Because such predictive instruments are meant to aid, not replace, clinical decisions, they can complement the often relied on gestalt approach to patient care, supporting (or refuting) physician beliefs regarding stabil- ity for outpatient management. An example of the potential utility of an AHFS predictive instrument was recently published by Hsieh and colleagues. They retrospectively analyzed an administrative database to derive and validate a predictive instrument that identified 19.2% of AHFS patients at low risk for 30-day adverse events.109 Their validated model incorporated vital signs, renal function, white blood cell count, and glucose as risk predictors. Events were infrequent in the low-risk cohort with inpatient mortality, in-hospital complication, and 30-day mortality rates of 0.7%, 1.7%, and 2.9%, respectively.
These results notwithstanding, a prospectively derived, mul- ticenter, ED risk stratification model for patients with signs and symptoms of HF is needed. Data suggest that emergency physicians would be comfortable discharging a patient if there was a combined overall risk of in-hospital events or 30-day mortality of 2%.220 Prospectively performed studies collecting ED-based data are needed to confirm preliminary findings and facilitate safe, early ED discharge. Such an approach, which is the focus of 2 ongoing National Heart, Lung, and Blood Institute grants being directed by emergency medicine investigators,223 has proven effective at safely decreasing admissions for low-risk patients with other disease processes such as acute coronary syndromes222–224 and community-acquired pneumonia.225–227 In- herent to this is the need to alter risk-stratification standards from prediction of remote adverse events (eg, 90 days, 1 year), which are highly dependent on subacute to chronic care and patient behavior, to those which occur sooner (eg, within 14 days) and are more likely to be associated with the patient’s acute HF
Similar to the use of repeat troponin
measurement or assessment of myocardial viability for acute coronary syndromes,224 incorporation of objective, evidence- based end points into evolving predictive instruments will provide important information regarding near-term risk that could, at last, be appropriately used in the acute setting to identify AHFS patients who are safe for early ED, observation unit, and hospital discharge.
ED Enrollment of Patients With AHFS It has become clear that there are many unanswered questions regarding ED care of the patient with AHFS. Evidence-based guidelines are needed for diagnostic, therapeutic, and disposition decision making. To conduct the clinical trials necessary to develop the foundation for an adequate evidence base, research- ers will have to enroll patients early in their AHFS presentation,
.org/ by guest on February 5, 2015