Albumin dialysis without anticoagulation in high-risk patients: an observational study.

Severe liver failure causes coagulopathy and high bleeding risk. Albumin dialysis with Molecular Adsorbent Recirculating System (MARS) (Gambro, Lund, Sweden) is useful for treatment. However, anticoagulation during its use is of uncertain value. We omitted heparin-saline priming and intradialytic heparin and examined its effects. Albumin dialysis was performed in critically ill patients with intermittent circuit saline flushes (2664 ± 2420 mL per treatment). A total of 12 patients (M : F = 10:2; age 49 ± 9 years) were thus treated: 6 for fulminant hepatic failure and 6 for acute-on-chronic liver failure. The overall hospitalization duration was 31 ± 30 days. A total of 44 treatment sessions were performed (average 8 ± 7 sessions per patient). Prescribed versus achieved MARS duration were 13 ± 3 versus 11 ± 4 h, P < 0.05. Twenty-three percent (10/44) of MARS sessions clotted, 11% (5/44) of treatments were electively terminated, and 2% (1/44) developed vascular catheter occlusion. Spontaneous bleeding occurred in 9% (4/44). Pre- versus post-MARS systemic and blood circuit transmembrane pressures (mm Hg), and albumin dialysate afferent and efferent pressures were all stable. Coagulation indices were (pre- vs. post-MARS): (i) prothrombin time (seconds): 36 ± 30 versus 42 ± 33, P = 0.143; (ii) activated partial thromboplastin time (seconds): 78 ± 43 versus 88 ± 45, P = 0.117; and (iii) platelet count (×10(3) /µL): 87 ± 40 versus 76 ± 48, P = 0.004. Systemic blood solute concentrations pre- versus post-MARS were: (i) serum urea (mg/dL): 22.4 ± 19.6 versus 14.0 ± 8.4, P < 0.05; (ii) serum creatinine (mg/dL): 2.8 ± 2.3 versus 1.9 ± 1.5, P < 0.05; (iii) total bilirubin (mg/dL): 29.5 ± 8.8 versus 20.5 ± 5.1, P < 0.05; and (iv) plasma ammonia (µg/dL): 186 ± 85 versus 129 ± 66, P < 0.05. Anticoagulant-free albumin dialysis remained effective despite frequent circuit clotting. This led to significant exacerbation of thrombocytopenia although bleeding risk remained low.

Real-time circuit pressures correlate poorly with circuit longevity in anticoagulant-free, predilution continuous venovenous hemofiltration.

BACKGROUND: Continuous venovenous hemofiltration (CVVH) in renal failure is compromised by circuit clotting. We hypothesized that adverse circuit pressures are predictive of clotting in circuits that last less than 24 h during predilution, anticoagulant-free CVVH. METHODS: This was a single-center retrospective study of 63 CVVH circuits of 13 critically ill intensive care unit patients with severe renal failure. Circuits were categorized into 'clotters' (C) or 'nonclotters' (NC), if spontaneous clotting occurred at <24 or ≥24 h from the start of CVVH, respectively. RESULTS: Effluent pressures and pre-filter pressures were more adverse in NC compared to C starting from 6 and 4 h before spontaneous clotting, respectively. Arterial pressures and return pressures were not significantly different in C versus NC. Blood flow rate settings, hemofiltration fluid replacement and effluent drainage rates in C versus NC were comparable. CONCLUSION: Real-time circuit pressure readings seem to offer only potentially limited prognostic value in predicting circuit clotting.