Acute renal failure is common in critically ill patients and necessitates renal replacement therapy in approximately 5% of the total intensive care population.1,2 Continuous venovenous hemofiltration (CVVH) is frequently used for that purpose, but the optimal infusion site for the replacement fluid remains controversial. In predilution mode, clearance may be limited because blood is diluted before it enters the filter; the latter, however, results in a lower blood viscosity, which may retard coagulation, prolong filter life, and thereby improve efficiency. Less clotting may thus render predilution preferable over postdilution when systemic anticoagulation is contraindicated.3 Hemodilution may also facilitate removal of urea from red blood cells.4–6 The argument against postdilution is that the increase in blood viscosity leads to early filter coagulation and to a shorter filter life, but clearance, for a given blood flow and ultrafiltrate rate, may be higher as compared with predilution in the absence of hemodilution.7 Clinical evidence favoring one or the other concept is scarce, however. Postdilution infusion was associated with a reduced filter life without any beneficial effects on azotemic control according to one study on patients with highly variable anticoagulant regimens,8 whereas filter life was longer on predilution in patients receiving intravenous nadroparin at the cost of lower creatinine clearances in another study.9 Lack of differences in filter life and clearance were also observed when nadroparin was used as a systemic anticoagulant.10 So, the available data concerning this issue are conflicting. Moreover, 31 and 9 included patients in the studies by van der Voort et al.9 and de Pont et al.,10 respectively, were rather small.
In this study, we evaluated a relatively large number of homogenous critically ill patients anticoagulated by heparin and filters used for CVVH in pre- and postdilution modes. We hypothesized equivalence of pre- and postdilution CVVH regarding filter life and azotemic control despite a difference in delivered dose.
Patients and Methods
This is a retrospective study. All patients admitted at the intensive care unit (ICU) of the Free University Medical Center between November 2004 and December 2006 treated by CVVH were studied. The ICU of the Free University Medical Center consists of two equally sized but separate units; one unit historically treated all patients with CVVH in the predilution mode and the other unit in the postdilution mode. In the past, both units were separate departments; in 2001, the units were merged with similar admission and discharge policies and rotating medical staff. The nursing personnel, however, remained fixed. The indication to start CVVH was based on clinical grounds. To warrant filter patency, patients were treated per protocol by heparin to reach an activated partial thromboplastin time (aPTT) between 55 and 65 seconds. In our unit, patients with an increased bleeding risk are treated with regional anticoagulation with citrate-containing replacement solution, and they do not form part of this study. Patients in the ICU for a minimum of 3 days and treated by heparin were only included in this study. Patients were excluded if only one filter was used with a filter life of <12 hours. CVVH was performed using a hemofiltration machine (Diapact, B. Braun, Melsungen, Germany). Vascular access was secured by inserting a 11-French double lumen catheter (GamCath, Gambro, Hechingen, Germany) into one of the three large veins (jugular, femoral, or subclavian). In all patients, a 1.9 m2 highly permeable cellulose triacetate hemofilter was used (Nipro UF205, Nissho Corporation, Osaka, Japan). Filters were routinely changed after 72 hours. Blood flow and flow of lactate- or bicarbonate-based replacement fluid were routinely set at 180 ml/min and 2 L/h, respectively, with net ultrafiltrate determined by treating physicians. CVVH was discontinued at the discretion of treating physicians, and patients were otherwise treated according to institutional guidelines. If not explicitly mentioned otherwise, the nurses were initially free to set the dilution mode in the way they were most comfortable with. Since December 2006, all patients are treated in predilution mode.
We designed a predefined checklist for this retrospective study. Our ICU has an electronic patient file where patients' details are stored. Baseline characteristics were retrieved, including age, gender, weight, height, prior chronic intermittent hemodialysis, and date and reason of admission. A severity of illness score at the time of ICU admission was generated by the Acute Physiology and Chronic Health Evaluation (APACHE) II.11 The Sequential Organ Failure Assessment (SOFA) score was evaluated at admission.12 The life of all filters used (up to number 12 per patient), the cause of filter termination, the prescribed hours of CVVH, downtime and last transmembranous pressure (TMP) were also recorded. Clotting as a reason of filter termination was defined as spontaneous clotting or a persistently high TMP (>200 mm Hg) prohibiting continuation of CVVH. Downtime was defined as the interval that CVVH was prescribed but not applied as a result of circuit clotting or as a result of transport to radiology or operation room. Other collected data included the aPTT, platelet count, and hematocrit (Ht), which were measured daily after the start of CVVH. Azotemic control was determined by evaluating the daily serum creatinine and blood urea concentrations on the first 3–10 CVVH days. The daily net ultrafiltrate volume, urine production, and fluid balance were registered. Delivered dose was defined as the ultrafiltration volume (substitution volume + net ultrafiltrate) delivered per kilogram preadmission body weight per hour; it was averaged per day and thus included downtime. For predilution, the ultrafiltration flow per hour (Quf) was adjusted by the following formula13,14:
where Qb = blood flow per minute and Qs = substitution flow per hour. Creatinine clearance is identical to the adjusted ultrafiltrate but is expressed as milliliters per minute. The filtration fraction (FF) was calculated. For postdilution, the following formula was used: FF = Quf/Qp × 100, where Qp = plasma flow [Qb × (1 − Ht)] per minute. For predilution, Qs was added to Qp [Quf/(Qp + Qs) × 100]. The length of stay in the ICU was recorded.
Values are summarized as mean ± SD or median ± interquartile range (IQR), in case of non-normal distribution (Kolmogorov-Smirnov test, p < 0.05). Values were logarithmically transformed in the latter case. The Mann-Whitney U test was used for univariate analyses. For categorical variables, the χ2 and Fisher's exact tests were used. Generalized estimating equations (GEE), taking into account the repeated measurements in the same patients, were used to evaluate determinants, including dilution mode, filter life, and azotemic control, until a maximum of 12 filters per patient, or per CVVH day, from day 0 (start of CVVH) to 2 or days 0–9. Generalized estimating equations are used to fit parameters of a generalized linear model where unknown correlation is present.15 The focus of GEE is on estimating the average response over the population rather than the regression parameters that would enable prediction of the effect of changing one or more covariates on a given individual. The course of logarithmically transformed creatinine and urea was compared among dilution modes after entering baseline values as covariates for adjustment. Filter life in the two groups is presented graphically as Kaplan-Meier survival curves; the log-rank test was used to compare filter life among the two groups. Exact p values are given and considered statistically significant if <0.05.
A total of 230 critically ill patients were treated by CVVH in the observation period. Sixty-three patients fulfilled the inclusion criteria, 36 patients in predilution and 27 patients in postdilution mode. A total of 243 filters were analyzed (up to 12 filters per patient), 132 in predilution and 111 in postdilution mode, so that between days 0 and 9, 123 filters were used in predilution and 97 in postdilution mode. There was no difference in baseline patient characteristics between pre- and postdilution modes (Table 1).
The filter life in pre and postdilution modes was comparable (Table 2, Figure 1). Furthermore, the filter life was inversely associated with TMP (p = 0.001). Table 3 describes similar anticoagulation by heparin among dilution modes. Clotting was the leading cause of filter termination (Table 4) and again predicted by the TMP (p = 0.007 by GEE) and not by pre- or postdilution mode, FF, or aPTT. In the predilution group, 19 (14%) filters were electively changed after 72 hours vs. 20 (18%) in the postdilution group. After exclusion of these filters, the filter life was still comparable between the two dilution modes (data not shown).
Delivered Dose and Azotemic Control
Table 2 describes lower delivered dose per day, expressed as ml/kg/h and adjusted for baseline values, in pre- versus postdilution. Figure 2 describes the daily course of creatinine and urea levels in the patients according to dilution mode, showing no differences. However, Table 5 shows that delivered dose had a direct effect on the fall of creatinine and urea levels in blood, between days 0–2 and 0–9.
Our study shows similar filter life for pre- and post-dilution CVVH in critically ill patients on heparin. Even though azotemic control depended on delivered dose, the 19% higher dose in postdilution than in predilution was insufficient to result in a difference in the course of serum creatinine and urea levels in our patients.
It is generally believed that hemodilution by prefilter- administered replacement fluid prolongs filter life in CVVH in studies on patients with highly variable and thus presumably suboptimal anticoagulant regimens.8,9 Our findings of a similar filter life when heparin is used systemically, however, confirm those by others in predilution (n = 8) and postdilution CVVH (n = 9) in patients all receiving anticoagulation by intravenous nadroparin.10 Apparently, the effect of heparin overwhelmed that of predilution in inhibiting filter coagulation. Moreover, it is debatable whether hemodilution has an inhibitory effect on coagulation, as a Ht <30% may more strongly activate coagulation when compared with a higher Ht.16 Furthermore, in an analysis on extracorporeal circuit thrombogenesis, no difference was found between pre- and postdilution regarding parameters on thrombin generation or platelet activation.10
The predilution mode carries the potential drawback of diluting blood before it enters the filter and thereby decreasing clearance, as compared with the postdilution mode, for a given blood flow and ultrafiltrate flow rate. In a small study on patients with and without anticoagulants, however, the azotemic control was comparable.8 One explanation is the loss of efficiency in the postdilution mode when filter life is shorter, leading to frequent filter changes and increased downtimes. In the aforementioned study comparing patients on systemic anticoagulation with nadroparin treated by either pre- or postdilution CVVH, no difference was observed in both filter life and azotemic control.10 A major drawback of the latter study, however, was its small sample size with only 15 filters being analyzed. Our larger study on 243 filters is nevertheless in line with these data. Indeed, we observed that azotemic control depended on delivered dose, but the hemodilution-associated fall in delivered dose with predilution, for a given flow rate, was apparently too small to result in a difference in azotemic control between the dilution modes. In contrast to prior suggestions,4–6 the data also argue against predilution facilitating urea clearance. As recent studies showed no survival benefit of delivered doses above 20 ml/kg/h or urea-targeted dialysis, it is doubtful whether the somewhat smaller clearance in the predilution mode is really a clinically important drawback.17–21
This study had several limitations. We have no detailed information concerning residual renal function. We used the urine production as derivative of residual function, which was comparable between the two CVVH modes. This was a retrospective single-center study with all inherent drawbacks and therefore our result should be interpreted with caution. The organization of the ICU, however, with two identical units except for the mode of CVVH applied, created a unique opportunity to compare pre- and postdilution modes, and patients were comparable among the units. To achieve complete uniformity between the ICU, all patients are treated in the predilution mode since December 2006. The choice for the predilution mode was predominantly based on our decision to adopt a regimen with regional anticoagulation with citrate-based replacement fluid in the predilution mode. The results of this study support that policy, as there are no clear advantages of postdilution CVVH.
In conclusion, we found, in retrospect, a similar filter life in pre- and postdilution CVVH in critically ill patients receiving systemic anticoagulation with heparin. Although delivered dose has a direct effect on the fall of urea and creatinine in blood, the hemodilution-induced 19% dose reduction in predilution CVVH is too small to result in a difference in azotemic control. Hence, predilution is equivalent to postdilution CVVH in heparinized patients.
The authors thank the staff of the intensive care and nephrology departments for their care of patients.
1.Uchino S: The epidemiology of acute renal failure in the world. Curr Opin Crit Care
12: 538–543, 2006.
2.Uchino S, Kellum JA, Bellomo R, et al
; Beginning and Ending Supportive Therapy for the Kidney (BEST Kidney) Investigators: Acute renal failure in critically ill patients: A multinational, multicenter study. JAMA
294: 813–818, 2005.
3.Tan HK, Baldwin I, Bellomo R: Continuous veno-venous hemofiltration without anticoagulation in high-risk patients. Intensive Care Med
26: 1652–1657, 2000.
4.Kaplan AA: Predilution versus postdilution for continuous arteriovenous hemofiltration. Trans Am Soc Artif Intern Organs
31: 28–32, 1985.
5.Cheung AK, Alford MF, Wilson MM, et al
: Urea movement across erythrocyte membrane during artificial kidney treatment. Kidney Int
23: 866–869, 1983.
6.Bellomo R, Ronco C: Continuous renal replacement therapy: Hemofiltration hemodiafiltration, or hemodialysis? in Ronco C, Bellomo R, Kellum JA (eds), Critical Care Nephrology
, 2nd ed. Philadelphia, PA, Saunders Elsevier, 2009, pp. 1354–1358.
7.Parakininkas D, Greenbaum LA: Comparison of solute clearance in three modes of continuous renal replacement therapy. Pediatr Crit Care Med
5: 269–274, 2004.
8.Uchino S, Fealy N, Baldwin I, et al
: Pre-dilution vs. post-dilution during continuous veno-venous hemofiltration: Impact on filter life and azotemic control. Nephron Clin Pract
94: c94–c98, 2003.
9.van der Voort PH, Gerritsen RT, Kuiper MA, et al
: Filter run time in CVVH: Pre- versus post-dilution and nadroparin versus regional heparin-protamine anticoagulation. Blood Purif
23: 175–180, 2005.
10.de Pont AJM, Bouman CSC, Bakhtiari K, et al
: Pre-dilution versus post-dilution during continuous venovenous hemofiltration: A comparison of circuit thrombogenesis. ASAIO J
52: 416–422, 2006.
11.Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med
13: 818–829, 1985.
12.Ferreira FL, Bota DP, Bross A, et al
: Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA
286: 1754–1758, 2001.
13.Marshall MR: Current status of dosing and quantification of acute renal replacement therapy. Part 2: Dosing paradigms and clinical implementation. Nephrology
11: 181–191, 2006.
14.Clark WR, Turk JE, Kraus MA, Gao D: Dose determinants in continuous renal replacement therapy. Artif Organs
27: 815–820, 2003.
15.Zeger SL, Liang KY, Albert PS: Models for longitudinal data: A generalized estimating equation approach. Biometrics
44: 1049–1060, 1988.
16.Stefanidis I, Heintz B, Frank D, et al
: Influence of hematocrit on hemostasis in continuous venovenous hemofiltration during acute renal failure. Kidney Int Suppl
17.Tolwani AJ, Campbell RC, Stofan BS, et al
: Standard versus high-dose CVVHDF for ICU-related acute renal failure. J Am Soc Nephrol
19: 1233–1238, 2008.
18.The VA/NIH Acute Renal Failure Trial Network, Palevsky PM, Zhang JH, O'Connor TZ, et al
: Intensity of renal support in critically ill patients with acute kidney injury. N Eng J Med
359: 7–20, 2008.
19.Vesconi S, Cruz DN, Fumagalli R, et al
; The DO-RE-MI Study Group: Delivered dose of renal replacement therapy and mortality in critically ill patients with acute kidney injury. Crit Care
13: R57, 2009.
20.The RENAL Replacement Therapy Study Investigators, Bellomo R, Cass A, Cole L, et al
: Intensity of continuous renal-replacement therapy in critically ill patients. New Eng J Med
361: 1627–1638, 2009.
21.Faulhaber-Walter R, Hafer C, Jahr N, et al
: The Hannover Dialysis Outcome study: Comparison of standard versus intensified extended dialysis for treatment of patients with acute kidney injury in the intensive care unit. Nephrol Dial Transplant
24: 2179–2186, 2009.