The study was performed in the intensive care unit (ICU) of the University Hospital Brussels and conducted in compliance with the Helsinki Declaration. The Central Ethical Committee of the University Hospital approved the study protocol (BUN 143201318818). Due to its retrospective and observational before-after design, the need for informed consent was waived.
Patients were eligible when presenting AKI requiring RCA-CVVH treatment. Inclusion criteria were: age ≥ 18 years, presence of at least an AKI “RIFLE injury” score [
16], and no contra-indication for RCA. Exclusion criteria were: patients already treated with CVVH during their ICU stay or receiving CVVH at time of enrolment, a high likelihood of dying within the first 24 h, impossibility to provide a correct vascular access, and Child-Pugh grade C liver cirrhosis. Baseline characteristics, including causes of AKI and incidence of shock are given in Table
1.
Table 1
Patient characteristics and causes of AKI
Number of patients (n) | 28 | 31 | |
Age (years) (median [range]) | 73 (65–82) | 68 (55–82) | 0.298 |
Male gender (n, %) | 17/28 (61 %) | 21/31 (68 %) | 0.598 |
Underlying disease/condition & cause of AKI (n, %) |
Sepsis | 9/28 (32 %) | 9/31 (29 %) | 0.98 |
Post cardiac surgery | 5/28 (18 %) | 10/31(32 %) | 0.33 |
Post general surgery | 7/28 (25 %) | 4/31(13 %) | 0.39 |
Other | 7/25 (25 %) | 8/31(26 %) | 0.90 |
APACHE II score (median [range]) | 32 (23–41) | 27 (20–34) | 0.109 |
Mechanical ventilation (days) (median [range]) | 11 (1–24) | 14.6 (2–26) | 0.167 |
Vasopressor requirement (n, %) | 86 % | 77 % | 0.451 |
Patients were not randomized to receive either PC10/2 or PC18 but consecutively included within one treatment group. In fact, RCA with PC10/2 was initially applied in all patients on CVVH. At a given moment, we decided to replace PC10/2 by PC18. The reasons for changing the concentration of citrate were the followings: initially, all patients requiring CVVH in our ICU received PC10/2 for regional anticoagulation. To obtain better metabolic control (and especially better control of metabolic acidosis), we decided to replace PC10/2 (10 mmol citrate/L) by PC18 (18 mmol citrate/L). For practical reasons (storage capacity, short shelf-life of the citrate liquids, potential prescription errors), the hospital pharmacy did not make the two citrate solutions simultaneously available but delivered the PC18 solution after the PC10/2 stock was entirely consumed. Thus, a first group of patients received PC10/2 and, subsequently, a second group of patients was started on PC18. Regarding citrate dosage used during the study, these two dosage regimens (PC 18 & PC 10/2) fall within accepted dosing range for citrate anticoagulation. Solutions were not homemade but are CE-labelled and marketed by Gambro-Baxter. In- and exclusion criteria were identical for both study periods
. CVVH was performed with the Prismaflex device (Gambro, Lund, Sweden) using an acrylonitrile 69 surface treated (AN69 ST) 150 membrane. Veno-venous access was obtained via a 13 F double-lumen polyurethane catheter (Joline, Swiss Confederation) inserted in the right internal jugular or a femoral vein. CVVH was delivered according to a dedicated protocol inspired by Tolwani et al. [
17] and presented in detail previously [
6]. This included standardized order sets and initial settings for all patients. Blood flow rate was set at 150 mL/min. Calcium chloride initially ran at 6 mL/h through a separate central venous line. Calcium infusion was titrated to maintain plasma ionized calcium levels between 1,0 and 1,2 mmol/L [
6]. PC10/2 was delivered before the filter and started at a rate of 2200 mL/h. A bicarbonate-buffered solution (Prismasol 2) was infused in post-dilution, starting at 800 mL/h. In the PC18 group, citrate was delivered at a starting rate of 1500 mL/h and another bicarbonate buffer (Prismocal B22) in post-dilution at 400 mL/h. Detailed characteristics of the citrate and substitution fluids are shown in Table
2. Arterial blood gases, lactate, and serum electrolytes, including systemic and post-filter ionized calcium, were analyzed every 4 h. Acid-base status was evaluated with the Stewart-Figge method. This approach postulates that acid-base balance and pH depend on the difference between concentrations of strong cations and strong anions (ie the strong ion difference; SID), the PaCO
2, and the total concentration of weak acids. It introduces the term “apparent strong ion difference” (SIDa) calculated as: ([Na
+] + [K
+] + [Mg
2+] + [Ca
2+]) - ([Cl
−] - [lactate
−]) (concentrations in mmol/L). The normal range for SIDa is approximately 40–44 mmol/L [
18]. Since this equation does not account for weak acids (albumin, phosphate and CO
2), the effective strong ion difference (SIDe) was calculated as (1000 × 2.46 × 10
−11 × pCO
2 / 10
-pH) + [albumin] × (0.12 × pH - 0.631) + [phosphate] × (0.309 × pH - 0.469) (with pCO
2 in mmHg, albumin in g/L and phosphate in mmol/L) [
17]. The SIDa to SIDe difference should equal zero unless unmeasured charges are present in the blood. These charges are captured by the strong ion gap (SIG = SIDa - SIDe). A positive SIG value represents unmeasured anions that are needed to account for the measured pH, measured levels of strong ion and weak acids, and to assure iso-electricity [
18]. Statistical analysis was performed using SPSS version 20 for Windows (SPSS Inc., Chicago, IL, USA). Chi-square and Fisher exact test were used to compare categorical variables between groups. The Mann-Whitney
U test was applied for comparison of non-normally distributed parameters and the Wilcoxon test was used for comparing variables within a group. Values were expressed as medians (range) unless indicated otherwise.
Table 2
Composition of citrate and bicarbonate-buffered solutions including their calculated SIDa
Citrate | 10 | 18 | | |
Citric acid | 2 | | | |
Sodium (Na+) | 136 | 140 | 140 | 140 |
Chloride (Cl-) | 106 | 86 | 111.5 | 120.5 |
Calcium (Ca2+) | | | 1.75 | 0 |
Magnesium (Mg2+) | | | 0.5 | 0.75 |
Lactate | | | 3 | 3 |
Hydrogen carbonate (HCO3-) | | | 32 | 22 |
Potassium (K+) | | | 2 | 4 |
Glucose | | | 6.1 | 6.1 |
SIDa | 30 | 54 | 29.75 | 21.25 |