Background
Methods
Patients
Materials and definitions
Study design
-
Phase I: the goal was to determine the target range of values for ECC ionized calcium at five minutes treatment (cCai t5’), avoiding TF. Given the short duration of TPE, we initially tested a low citrate concentration aiming a relatively high cCai t5’ of 0.35-0.40 mmol/L. This approach was chosen to minimize the patient exposition to citrate given the convective nature of TPE. Treatment efficacy was characterized by the absence of TF or PTF. The target range for cCai t5’ (corresponding to RCA intensity) should enable treatment efficacy in more than 80% of treatments. In order to assess the determinants of the response to citrate, detailed information were collected.
-
Phase II: we tested an empirical approach of the citrate concentration which is required in the ECC to reach the target range of values for cCait5’, and proposed a preliminary protocol that would allow to obtain a maximal number of treatments with cCai t5’ in the target therapeutic range.
-
Phase III: Based on data collected during the first and second phases, a mathematical modeling approach has been used to develop the final treatment protocol.
-
Phase IV: we assessed clinically the efficacy and safety of the new treatment model.
-
Treatment monitoring included serial measurements of characteristic biological parameters at key time points (5, 30, 60, 90 min, end of treatment). This timing was based on the pharmacokinetics of citrate as reported in the medical literature [9]. Additional measurements were performed if deemed necessary by the clinician. Changes in citrate infusion rate were correlated to observed changes in cCai and other biological variables in order to assess the response to citrate. Treatment efficacy was again characterized by the absence of TF or PTF.
Statistical analysis
Results
Patients
Characteristics of the patients | |
---|---|
Number of patients | 14 |
Sex, n, Female/Male | 5/9 |
Age, years (mean ± standard deviation) | 52.6 ± 17.83 |
Treatment characteristics | |
Median number of treatments per patient | 2.5 |
TPE indication (N = 47), n (%) | |
Thrombotic microangiopathy | 25 (53) |
Acute humoral rejection | 8 (17) |
ANCA-associated vasculitis | 7 (15) |
Antiglomerular basal membrane vasculitis | 3 (6) |
Cryoglobulinemia associated vasculitis | 2 (4) |
Neurological diseases | 2 (4) |
TPE during active bleeding, n (%) | 8 (17) |
Treatment characteristics
-
Phase I: determination of the target range values for cCai t5’, avoiding TF.A standardized prescription of citrate and calcium was tested in the first five TPEs performed in three patients, Citrate was perfused pre-filter at a concentration of 2.4 – 3.0 mmol/L of whole blood, aiming a target cCai of 0.35-0.40 mmol/L which correspond to a relatively low concentration as mentioned earlier.This infusion rate led to the predefined cCai target, but was associated with TMP elevation and TF (Table 2). For the next treatments, we therefore lowered the cCai target range to 0.25-0.35 mmol/L, which corresponds to the target mostly reported in the literature in various modalities of RRT [6, 10, 11]. This range, deemed appropriate, has been tested in the next successive treatments.Table 2First 5 TPE aiming a cCait5’ of 0.35-0.40 mmol/LN° Pt. TttHt (%)pCai t0’CiWBcCai t5’cCai maxPTFTFA.1401.142.40.400.42YNB.1271.002.40.370.37YYB.2251.072.40.410.41YNB.3260.832.40.400.47YYC.1271.273.00.390.39yY
-
Phase II: Empirical approach of the citrate concentration needed in the ECCFor the next treatments, we empirically increased WBCi to 3.6 mmol/L in order to reach the new target range for cCai t5’ (ie 0.25-0.35 mmol/L). Based on physiological considerations and an analysis of the first treatments parameters, pre-treatment pCai and the Ht were strongly associated with the efficacy of a given dose of citrate. Indeed, since citrate is prescribed in whole blood according to the blood pump and does not enter the erythrocyte, its plasma concentration is assumed to depend on the patient's Ht (Fig. 2, equation 2).Concerning pCai, we observed that the citrate concentration required in plasma to lower ionized calcium concentration by one millimole per liter (equivalent to the ratio pCi/pCait0’), was about 4.6 millimoles. Based on the assumption that this value is constant, we developed a first prescription protocol for RCA, which integrated cCait0’ and Ht (Fig. 2, equation 3 and 3’).The objective of this preliminary protocol was to obtain a large number of treatments in the target therapeutic range for cCai t5’, that would allow later the development of a mathematical modeling.Thus, this preliminary protocol or, alternatively a fixed citrate dosing (WBCi) of 3.6 mmol/L was applied in the next 42 following treatments beyond the first step. Among these, 38 were sufficiently well documented for the plasma values of both pre-treatment ionized calcium and circuit ionized calcium. These 38 TPE could be therefore included in the next phase (modeling process).In these 38 TPE, no TF was observed. Aside from two cases (2/38, 5.3%) of PTF that occurred in relation with technical problems (catheter dysfunction, interruption of RCA), any rise of the TMP (4/38, 10.5%) was systematically associated with at least one measurement of cCai above or equal to 0.35 mmol/L at any time during the treatment. Therefore, we adjusted the target range of values for cCait5’ from 0.25-0.35 to 0.24-0.33 mmol/L.
-
Phase III: Modeling approachWhen circuit plasma ionized calcium concentration is within the target range (0.24-0.33 mmol/L, in 30 out of 38 treatments), the ratio pCi/pCait0’ expresses the amount of citrate (in mmol) in one liter of plasma that is required to neutralize one millimole of calcium. As mentioned previously, we assumed that pCi/pCait0’ was constant (4.6 mmol/mmol) and thus may be used for the calculation of whole blood citrate dosage in any individual patient, knowing pre-treatment pCait0’ and Ht (Fig. 2, equation 3’).We further analysed pCi/pCait0’ and the ratio pCi/DeltaCai, where DeltaCai is the difference between pCait0’ and cCai t5’. DeltaCai also represents the amount of calcium becoming chelated by citrate in 1 l of plasma.For relatively higher pre-treatment pCai, we found that the ratio pCi/DeltaCai is lower (Fig. 3a, b, c). Indeed, less citrate in regard to calcium is needed to reach the therapeutic target. This may be explained by a larger availability of plasma calcium for citrate in case of high plasma calcium concentration. Accordingly, pre-treatment hypocalcemia was associated with a relative resistance to citrate treatment.×Using a polynomial regression model, we showed a highly significant correlation between the ratio pCi/deltaCai(t0’-5’) and pCait0’, in a given individual and for all the treatments taken together (correlation coefficient R2 = 0,723, Fig. 3b and c). A similar correlation was noted between massic flow rates of both citrate and calcium (Fig. 3a). This correlation allowed to progress in the development of a prescription protocol (equation 4, Fig. 2).In order to obtain an expression of the citrate concentration pCi (corresponding to the prescription of citrate in plasma) from pCait0’, we introduced empirically in equation 4 (Fig. 2) a value for cCai t5’ in the middle of the target range, ie 0.285 mmol/L (equation 5, Fig. 2, Fig. 4).×Equation 5’, illustrated in Fig. 2, integrates the Ht and may be used for the determination of the initial dose of citrate in whole blood.
-
Phase IV: Clinical safety and efficacy analysisThe retrospective application of this new formula to the values of pre-treatment pCai of the 47 treatments of the study, allowed to improve both citrate underdosing (cCai at t5' > 0.33 mmol/L in 4 treatments) and overdosing (cCai at t5' <0.24 mmol/L in 4 treatments). This was at the price of a slight rise in citrate dosing in 6 treatments during which dosing was already adequate.Concerning safety, we managed to adapt calcium infusion according to a pre-defined protocol. A cautious clinical monitoring and the timing of the serial measurements of pCai allowed for the identification of biological anomalies which were corrected with the application of a systematic algorithm (Additional file 1: Table S1 and Additional file 2: Table S2). Clinically significant hypocalcemia was observed in four (8.5%) treatments and were corrected with calcium chloride administration without complication. Total calcium rose significantly during treatment as a consequence of calcium administration to maintain pCai balance (Fig. 5). This was associated with a concurrent elevation of the ratio pCa total/pCai (Table 3). As expected, the highest values of the pCa total/pCai ratio were observed during TPE with large amount of FFP as the substitution solution. Unexpectedly however, there were no proportional differences in the amount of the calcium infused nor in the concentration of post-treatment pCai (Fig. 6).Table 3Pre-treatment and post-treatment plasma values of pCai, pCa total, the ratio pCa total/pCai, pH and bicarbonates (mean ± SD))Pre-treatment (t0’)Post-treatmentpCai (mmol/l)1.00 ± 0.11.02 ± 0.07pCa total (mmol/l)2.04 ± 0.142.58 ± 0.25pCa total/pCai-2.53 ± 0.28pH7.47 ± 0.057.48 ± 0.04Bicarbonate (mmol/l)26.4 ± 4.927.6 ± 5.5××We did not observe any new episode of bleeding, worsening of a pre-existing bleeding, nor clinically significant electrolytes and acid–base disturbances (Additional file 3). We considered as a clinically significant acid–base disturbance any metabolic alcalosis or acidosis that was complicated by an adverse clinical event requiring a medical intervention. Mean bicarbonate elevation per treatment appeared to be proportional to FFP volume, as was pH elevation (Fig. 7). In four patients undergoing 5 to 6 daily consecutive TPE using FFP for more than 80% of total replacement solutions (mean total FFP = 21 L ± 3.71 SD), mean bicarbonate variation throughout the period of treatment was significant (+14.5 mmol/L ± 8.76 SD, from an initial mean bicarbonate concentration of 20 ± 4.32 SD mmol/L). However, mean pH variation was relatively modest (+0.1 ± SD 0.06 units). None of them required medical intervention.×