Management of Overhydration in Heart Failure Patients

Ultrafiltration

Isolated venovenous UF is a method of decongestion that can be used as an alternative to loop diuretics. UF has been made feasible with the advent of simplified devices that permit volume removal with peripheral venous access, adjustable blood flow, and small extracorporeal blood volumes ( Fig. 119.1 ). With this therapy, plasma water is produced from whole blood across a semipermeable membrane (hemofilter) in response to a transmembrane pressure gradient that is driven by hydrostatic forces generated by extracorporeal pumps. These hydrostatic forces can be adjusted manually, allowing for tightly controlled UF fluid removal rates. The solute concentration in the ultrafiltrate is equal to that in the water component of the plasma, allowing for effective isotonic sodium removal from the patient. In 2002 the Aquadex System 100 peripheral venovenous system (Gambro UF Solutions, Minneapolis, MN) was approved by the US Food and Drug Administration (FDA) for clinical use based on the results of the Simple Access Fluid Extraction (SAFE) trial. This study showed that, in 21 congested ADHF patients, the removal of an average of 2600 mL of ultrafiltrate during an 8-hour treatment period resulted in a mean weight loss of approximately 3 kg without changes in heart rate, blood pressure, SCr, electrolytes, or the occurrence of major adverse events. Several UF studies conducted thereafter were discussed at the 11th Acute Dialysis Quality Initiative (ADQI) meeting and are presented in the following section.

Pilot Studies

One pilot study looked at whether UF begun within 12 hours of admission could restore euvolemia safely, permit discharge in 3 days or less, and prevent 90-day rehospitalization in 20 ADHF patients with diuretic resistance (defined as SCr ≥ 1.5 mg/dL combined with daily oral furosemide doses ≥ 80 mg or equivalent doses of other loop diuretics). Vasoactive drugs and more than one dose of intravenous (IV) loop diuretic were prohibited before initiation of UF. An average of 8654 ± 4205 mL was removed with 2.6 ± 1.2 8-hour UF sessions. Twelve patients (60%) were discharged in 3 days or less. Improvement in weight ( p = .006), Minnesota Living with Heart Failure scores ( p = .003), and Global Assessment ( p = .00003) observed after UF persisted at 30 and 90 days. Levels of B-type natriuretic peptide (BNP) were decreased after UF (from 1236 ± 747 pg/mL to 988 ± 847 pg/mL) and at 30 days (816 ± 494 pg/mL) ( p = .03). Blood pressure, renal function, and medications were unchanged. Remarkably, in seven patients with hyponatremia (serum sodium ≤ 135 mg/dL), sodium increased from pretreatment values at discharge ( p = .042) and at 90 days ( p = .017). Given that ultrafiltrate is isotonic with plasma, the rise in serum sodium was not attributed to direct effects of UF, but rather to attenuation of neurohormonal activation, as indicated by the decrease in plasma BNP levels without worsening renal function. The results of this pilot study suggested that, in ADHF patients with diuretic resistance, UF initiated early before therapy with IV loop diuretics effectively decreased readmissions and length of stay, with clinical benefits extending to 90 days. This preliminary study has important limitations, including a small sample size, lack of a control group, and the now-obsolete FDA-mandated restriction of each UF course to 8 hours. Nevertheless, the observed benefits may reflect the fact that fluid removal by UF occurred before upregulation of neurohormonal activity by IV loop diuretics. In the Relief of Acutely Fluid-Overloaded Patients with Decompensated Congestive Heart Failure (RAPID-CHF) trial, 40 patients were randomized to either a single 8-hour course of UF at fluid removal rates determined by the treating physician plus usual care, or to usual care alone. Weight loss, the primary end point of the study, failed to reach statistical significance ( p = .240). However, compared with the usual care group, UF-treated patients had greater net fluid loss at 24 hours (4650 mL vs. 2838 mL; p = .001) and at 48 hours (8145 mL vs. 5375 mL; p = .012), and greater 48-hour improvement in dyspnea ( p = .039) and other heart failure symptoms ( p = .023). Usual care and UF were similar in terms of renal function, electrolytes, heart rate, systolic blood pressure, and duration of the index hospitalization. As in the previous study, effective decongestion and clinical improvement were observed with early initiation of UF, before elevation of SCr levels resulting from IV loop diuretic dosing. Again, study limitations must be recognized, given the small sample size and lack of assessment of outcomes beyond 48 hours. A single-center study of 11 patients with advanced diuretic resistant heart failure (defined by pretreatment average SCr of 2.2 mg/dL, mean estimated glomerular filtration rate [eGFR] of 38 mL/min [with 6/11 having eGFR < 30 mL/min]), mean daily IV furosemide dose of 258 mg, including nine patients [82%] with documented severe right ventricular [RV] dysfunction, and three patients [27%] with pericardial constriction) provides valuable insight on the appropriate use of UF in ADHF patients. This study sought to remove a goal of 4 L of fluid with each 8-hour UF course, which was achieved in only 13 of 32 treatments (41%). Five patients (45%) experienced an increase in SCr of more than 0.3 mg/dL, and five patients required hemodialysis (HD). There was no obvious correlation between amounts of fluid removed by UF and the need for HD. The severity of illness of these patients is exemplified by the 55% 6-month mortality noted in this trial. Such mortality rate was identical to that occurring in the medical therapy arm of the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial, and it exceeds the 6-month mortality ever reported in any other heart failure clinical trial.

Based on these observations, one can deduce several key points in the use of UF therapies: (1) isolated venovenous UF does not significantly alter the outcomes of patients with end-stage heart failure, and (2) fast fluid removal rates should be used with caution because they can be very detrimental to cardiorenal physiology, particularly in patients with RV dysfunction who are exquisitely susceptible to intravascular hypovolemia because of the storage of a larger proportion of blood in the venous circulation. Thus rapid removal of volume with aggressive UF in heart failure patients with RV dysfunction can decrease renal perfusion pressure, cause a rise in SCr, and convert nonoliguric renal dysfunction into oliguric failure and subsequent dialysis dependence. High doses of IV loop diuretics before UF, by intensifying neurohormonal activation, may predispose the kidney to injury by additional fluid removal with UF.

Ultrafiltration Versus Intravenous Diuretics for Patients Hospitalized for Acute Decompensated Heart Failure Trial

The goal of the Ultrafiltration Versus Intravenous Diuretics for Patients Hospitalized for Acute Decompensated Heart Failure (UNLOAD) trial was to compare the safety and efficacy of an early strategy of UF versus standard IV diuretic therapy in ADHF patients with two or more easily detectable signs of congestion. To achieve this goal, randomization had to occur within 24 hours of hospital admission and a maximum of two IV loop diuretic doses were permitted before randomization. A total of 200 patients (aged 63 ± 15 years; 69% men; 71% ejection fraction ≤ 40%) were randomized to UF or IV diuretics. At 48 hours, weight (5.0 ± 3.1 kg vs. 3.1 ± 3.5 kg; p = .001) and net fluid loss (4.6 L vs. 3.3 L; p = .001) were greater in the UF group. Dyspnea was improved similarly in the two groups. At 90 days, the UF group had fewer patients rehospitalized for heart failure (18% vs. 32%; p = .037) and fewer unscheduled visits for worsening heart failure (21% vs. 44%; p = .009). A similar percentage of patients with increases in SCr levels exceeding 0.3 mg/dL was noted in the UF and standard care group at 24 hours (14.4% vs. 7.7%; p = .528), at 48 hours (26.5% vs. 20.3%; p = .430), and throughout the 90-day follow-up period. Occurrences of hypokalemia (serum potassium < 3.5 mEq/L) were fewer in the UF than in the diuretic group (1% vs. 12%; p = .018), and episodes of hypotension during treatment were rare in both groups (4% vs. 3%). Complications related to UF included clotting of five filters, one catheter infection, and the requirement for HD in one patient deemed to have congestion refractory to UF.

The UNLOAD trial lacked treatment targets, blood volume assessments, and cost analysis. However, the salient findings of this trial still provide valuable lessons: an early strategy of UF, initiated before the administration of high-dose IV diuretics, effectively reduces congestion and 90-day heart failure–related rehospitalizations in ADHF patients. A posthoc analysis from the UNLOAD trial reviewed the outcomes of 100 patients treated with UF compared with those of 100 control group subjects divided according to whether they had received IV diuretics by continuous infusion (n = 32) or bolus injections (n = 68). Despite similar amounts of fluid removed by UF and continuous IV diuretic infusion, at 90 days heart failure–related rehospitalizations plus unscheduled visits (rehospitalization equivalents) were fewer in the UF group than in continuous IV diuretic infusion group ( p = .016). Volume overload in heart failure (HF) patients occur in relation to an increase and abnormal distribution of total body sodium. The greater reduction of total body sodium by isotonic fluid removal may be more effective than elimination of hypotonic fluid by diuretics or isolated free water by vasopressin V ₂ receptor blockers. It is also possible that prehospitalization diuretic use reduces the natriuresis achievable with the subsequent administration of IV loop diuretics. Increased central venous pressure (CVP) is associated independently with worsening renal function. The increased amounts of sodium and water reabsorbed by the kidney because of neurohormonal upregulation predominantly fill the compliant venous circulation, increasing CVP. Transmission of venous congestion to the renal veins further impairs GFR. Successful lowering of CVP without development of WRF calls for effective use of UF with establishment of fluid removal rates that do not exceed capillary refill rates, so adequate intravascular volume is maintained. This raises the concept of the plasma refill rate (PRR) (mL/min), which is a measurement of the fluid volume transport from the interstitium into the vascular space during UF, expressed as filtrate volume/time, where time is the duration of UF. Fluid removal rates can be titrated to be equal or lower than the PRR so that refilling of the intravascular space by the excess fluid in the interstitial space is maintained, and the likelihood of worsening intravascular volume depletion is reduced. Maintenance of an adequate blood volume reduces the risk of the development of WRF during decongestion of ADHF patients.