Basel, Switzerland - Levels of C-reactive protein (CRP) can independently predict two-year mortality in patients presenting to the emergency room with acute heart failure, a novel finding that could expand the inflammation biomarker's prognostic value and raises questions about the syndrome's pathophysiology, according to data from a previously published trial [1].

The finding could "encourage the use of this simple, widely available, reliable, and inexpensive parameter in the risk stratification of patients with acute HF," write Dr Christian Mueller and associates (University of Basel, Switzerland) in the April 2006 issue of the American Heart Journal. "Patients with a CRP level >25 mg/L should receive particular attention, as they do have significantly increased short- and long-term mortality."

Elevated CRP may reflect a greater burden of proinflammatory cytokines that acts as a trigger for decompensation.

The analysis raises the possibility that inflammation promotes "a distinct subtype of acute heart failure that may be unrelated to the existence of chronic heart failure," write Dr G Michael Felker and Dr Gadi Cotter (Duke Clinical Research Institute, Durham, NC) in an accompanying editorial [2]. "Elevated CRP may reflect a greater burden of proinflammatory cytokines that acts as a trigger for decompensation and also plays a role in the long-term progression of heart failure."

The 217 patients in the analysis had been confirmed to have acute HF in the Brain Natriuretic Peptide for Acute Shortness of Breath Evaluation (BASEL) study, a prospective single-blind comparison of brain natriuretic peptide (BNP) levels vs then-standard criteria for evaluating 452 patients with dyspnea in the emergency department. Its primary findings, that BNP assessment can accelerate correct acute-HF diagnoses and reduce hospitalization time and costs, were published in 2004 [3].

Patients with admission CRP levels in the highest tertile showed a 60% higher two-year mortality, the primary end point, compared with those in the lowest CRP tertile. The highest-tertile patients also went to the intensive-care unit and died in the hospital significantly more often and had a higher composite two-year rate of death or rehospitalization.

Other findings:

• Cumulative mortality curves by CRP level diverged throughout the first three months but seemed parallel thereafter, suggesting that the biomarker may predict predominantly short-term outcome, the group writes.
• CRP remained an independent mortality predictor—HR 1.4 (95% CI 1.1-1.8) for each tertile increase—after adjustment for age, sex, CAD, chronic obstructive pulmonary disease, renal function, and troponin and BNP levels.
• CRP's predictive significance persisted even after exclusion of patients with pneumonia or a history of fever at admission and after controlling for LVEF (p=0.016).
• CRP elevations were more pronounced (p=0.022) among the patients with LVEF >45% compared with those with lower measurements.

Some evidence suggests that acute HF can trigger a systemic anti-inflammatory response, but the reverse may also be true: inflammation might precede and promote acute HF, which, if true, suggests a potential role for anti-inflammatory therapies in the acute syndrome, according to Mueller and associates.

Measurement of CRP and troponin may be more useful at the time of hospitalization, whereas BNP levels during and at the completion of hospitalization may reflect adequacy of treatment and allow appropriate intensity of follow-up.

Elaborating on that point, Felker and Cotter speculate whether interventions against the generalized immune response or some of its components may be more appropriate in acute rather than chronic HF. That, they write, would be consistent with the failure of treatments aimed at suppressing tumor necrosis factor (TNF) in patients with chronic systolic HF, animal evidence for an association between inflammatory-mediator elevations and physiologic responses mirroring signs of acute HF, and evidence that acute TNF or interleukin-6 infusions may promote diastolic dysfunction. This last could potentially be related to Mueller et al's findings of higher CRP elevations in patients with HF and preserved LV function even as the marker's prognostic value was independent of LVEF.

They speculate that BNP's strength as a mortality predictor in acute HF, which seems to be better at discharge than admission, may relate to the success or failure of acute therapy or to the severity of underlying chronic HF. "Theoretically," write Felker and Cotter, "measurement of CRP and troponin . . . may be more useful at the time of hospitalization, whereas BNP levels during and at the completion of hospitalization may reflect adequacy of treatment and allow appropriate intensity of follow-up."