Diuretic Resistance Predicts Mortality
Background: In patients with chronic heart failure (CHF), diuretic requirements increase as the disease progresses. Because diuretic resistance can be overcome with escalating doses, the evaluation of CHF severity and prognosis may be incomplete without considering the intensity of therapy.
Methods: The prognostic importance of diuretic resistance (as evidenced by a high-dose requirement) was retrospectively evaluated in 1153 patients with advanced CHF who were enrolled in the Prospective Randomized Amlodipine Survival Evaluation (PRAISE). The relation of loop diuretic and angiotensin-converting enzyme inhibitor doses (defined by their median values) and other baseline characteristics to total and cause-specific mortality was determined by proportion hazards regression.
Results: High diuretic doses were independently associated with mortality, sudden death, and pump failure death (adjusted hazard ratios [HRs] 1.37 [P = .004], 1.39 [P = .042], and 1.51 [P = .034], respectively). Use of metolazone was an independent predictor of total mortality (adjusted HR = 1.37, P = .016) but not of cause-specific mortality. Low angiotensin-converting enzyme inhibitor dose was an independent predictor of pump failure death (adjusted HR = 2.21, P = .0005). Unadjusted mortality risks of congestion and its treatment were additive and comparable to those of established risk factors.
Conclusions: The independent association of high diuretic doses with mortality suggests that diuretic resistance should be considered an indicator of prognosis in patients with chronic CHF. These retrospective observations do not establish harm or rule out a long-term benefit of diuretics in CHF, because selection bias may entirely explain the relation of prescribed therapy to death.

In patients with chronic congestive heart failure (CHF), the assessment of disease severity and estimation of mortality risk are important for clinical management and in triage for special interventions such as transplantation. The prognostic value of various clinical and laboratory parameters has been demonstrated in numerous CHF studies. In practice, congestive symptoms and signs, left ventricular (LV) dysfunction, reduced exercise capacity, hypotension, azotemia, and arrhythmias are the indicators most commonly used to guide treatment. However, prognostic algorithms have not attempted to relate the severity of CHF to the intensity of therapy.

Medical therapy for CHF must be intensified as the disease progresses, and this is particularly true of diuretics. Clinicians assume that patients who require higher doses of diuretics to prevent fluid retention and control symptoms are sicker than those who require less aggressive therapy, but the prognostic importance of such factors has not been examined or compared with that of established risk markers. Despite the subjective nature of clinical decision-making, we hypothesized that consideration of the degree of congestion together with its treatment would be prognostically relevant in patients with chronic CHF. Therefore, we retrospectively evaluated the relation of medication doses and other baseline variables to mortality and mode of death in patients with advanced CHF who were enrolled in the Prospective Randomized Amlodipine Survival Evaluation (PRAISE).