Effects of Initiation of Continuous Renal Replacement Therapy on Hemodynamics in a Pediatric Animal Model

Abstract

There are no studies analyzing the initial hemodynamic impact of continuous renal replacement therapy (CRRT) in children. We have performed a prospective observational study in 34 immature Maryland pigs to analyze the initial hemodynamic changes during venovenous CRRT. The heart rate, blood pressure, central venous pressure (CVP), pulmonary arterial occlusion pressure (PAOP), pulmonary capillary wedge pressure, temperature, and cardiac output (CO), simultaneously by pulmonary arterial thermodilution and femoral arterial thermodilution, were measured at 30-min intervals during 2 h. Venovenous CRRT induced an initial significant diminution of volemic hemodynamic parameters (intrathoracic blood volume, global end-diastolic volume, stroke volume index, PAOP, and CVP). Simultaneously, a significant increase in systemic vascular resistance index and left ventricular contractility, and a decrease in CO, was observed. We conclude that CRRT in a pediatric animal model induces initial hypovolemia, and a systemic cardiovascular response with vasoconstriction and increase in ventricular contractility.

Keywords:

• continuous renal replacement therapy
• children
• hemofiltration
• hemodynamic
• cardiac output
• acute renal failure

INTRODUCTION

More recently, treatment of acute renal failure (ARF) in children has changed considerably. Venovenous continuous renal replacement therapy (CRRT) have become the technique of choice in the treatment of ARF in critically ill children, allowing continuous fluid removal with better hemodynamic tolerability than peritoneal dialysis and hemodialysis.[1–3] Several studies have analyzed the impact of renal replacement therapies on cardiovascular system in adults.[4–7] The hemodynamic repercussion of the connection to an extravascular circuit and the negative fluid balance is more relevant in children than in adults. However, there are no studies analyzing the hemodynamic impact of CRRT in children. This study analyzes the hemodynamic response to CRRT in a pediatric animal model.

MATERIAL AND METHODS

The study was carried out on 34 immature Maryland miniature pigs, isogenic for three loci on the major histocompatibility complex, weighing 8 to 16 kg, mean 10 (SD 1.6). The study was approved by the Institutional Review Board for the care of animal subjects, and the care and handling of the animals were in strict accordance with the guidelines for ethical animal research. The animals were initially premedicated with ketamine 15 mg/kg and atropine 0.02 mg/kg and, after endotracheal intubation, were connected to a ventilator (Boyle Modular, BOC) in volume control mode. During CRRT, anesthesia was maintained with 2% halothane and propofol in continuous infusion, and when it was necessary, we administered bolus of fentanyl for analgesia and pancuronium bromide for neuromuscular blockade. Ventilation was adjusted to maintain the Pao₂ above 13.3 kPa and the Paco₂ between 4.7 and 5.3 kPa. The femoral vein and the right and the left internal jugular veins were exposed by cutdown and cannulated. A 5.5- or 7.5-French Swan-Ganz catheter (Baxter, Healthcare Corporation, Edwards, Irvine, CA, USA), with a distance of 15 cm from the atrial port to the thermistor and two 18-gauge catheters to perform CRRT were introduced. A 5-French Thermodilution Pulsiocath arterial catheter (Pulsion Medical Systems, Munich, Germany) was inserted into the femoral artery for continuous cardiac output and blood pressure monitoring, and for blood gas determinations and analyzes.

A vacuum-driven, tubular, blood pumping device in 19 animals[8] or an IVAC 571 volumetric pump infusion (IVAC Corporation, San Diego, CA, USA) in 15 animals were used. We have previously demonstrated that both pumps are adequate to CRRT in an animal pediatric model.[9],[10] We used a pediatric 0.2 m² polysulphone hemofilter FH22 (Gambro Health care, Lakewood, CO, USA). The priming volume for the circuit and the hemofilter was 41 mL with volumetric pump and 48 mL with tubular pump.[8],[9] The circuit and filter were rinsed with 2 L of normal saline containing 5.000 IU/L heparin. At the beginning of the CCRT, the venous line was connected to the inflow port of the circuit without discarding the rising fluid. Three different blood flows were maintained at 5 (11 experiments), 15 (11 experiments), and 30 mL/min (10 experiments). The dialysate was pumped through the filtrate compartment at 100 mL/h. Pediatric urinary chambers (urinometers) were used for collection and volume measurement of the ultrafiltrate. The ultrafiltrate flow rate was continuously measured, but it was not regulated. Mean ultrafiltrate flow was 235 mL/h. Replacement fluid (Ringer's lactate solution) was delivered into the postfilter limb via volumetric pumps (IVAC 565 and 591) continuously adjusted continuously adjusted at intervals of 5 to 10 min at a similar rate to the ultrafiltrate flow. Mean replacement fluid administered was 222 mL/h. Ringer's solution was used for the dialysis fluids that were delivered at 100 mL/h in 17 experiments. Animals did not receive any vasoactive drug during the experiment.

Heart rate, blood pressure, central venous pressure (CVP), pulmonary arterial occlusion pressure (PAOP), and arterial temperature were measured continuously. Cardiac output (CO) was measured in each animal before and after circuit connection (in the first 2 to 5 min), and at 30-min intervals for 2 h. For each determination, 5 mL of iced 0.9% saline, at a temperature below 8°C, was manually injected through the atrial port of the pulmonary artery catheter and the CO was measured simultaneously using the Swan-Ganz catheter and the PiCCO monitor. The monitor for pulmonary arterial thermodilution method was the Kolormon GC/FE 7276 (Kontron, Inc. Watford, UK). The femoral artery catheter was connected to a pulse contour analysis computer (PiCCO, Pulsion Medical Systems) for measurements by the femoral arterial thermodilution technique. Two determinations were made on each occasion, and the mean of these two results was recorded. The following hemodynamic parameters were registered every 30 min: CO, PAOP, global end-diastolic volume index (GEDVI), intrathoracic blood volume index (ITBVI), extravascular lung water index (EVLWI), stroke volume index (SVI), systolic volume variation (SVV), cardiac function index (CFI), left ventricular contractility (Dp/dtmax), and systemic vascular resistance index (SVRI). Arterial blood gases, hematocrit, sodium, potassium, and calcium were determined before the beginning of CRRT and every 30 min.

Statistical tests analysis of variance (ANOVA) was used to compare the values between the two techniques and the three blood flows. ANOVA for repeated measures and Kruskal-Wallis test were used to compare the evolution of the values. Differences were considered significant at p < 0.05.

RESULTS

At the beginning of the CRRT, a significant decrease of volemia-related hemodynamic parameters, ITBVI (Figure 1), GEDVI CVP, and PAOP was observed (p < 0.05) (Table 1). The SVV also increased, but differences did not reach statistical significance. Simultaneously, an increase in SVRI (Figure 2) and left ventricular contractility (Dp/dtmax) (Figure 3) occurred, and CO decreased (Figure 4) (p < 0.05). PAOP increased slightly, and there were no initial significant changes in heart rate and arterial pressure (Table 1). These hemodynamic changes progressively increased during the 2 h of CRRT, although more gradually than at the beginning (Table 1). At 2 hours, the increase in heart rate and arterial pressure was significantly higher than at the beginning. There were no significant changes in neither CFI nor EVLWI. Arterial temperature progressively decreased during the experiment (p < 0.05) (Table 1). There were no significant differences in temperature changes between the three different blood flows.

Figure 1. Evolution of intrathoracic blood volume index (ITBVI) (mL/m²) (p < 0.05) (analyzis of variance [ANOVA] for repeated measures). *p: statistical comparison with values before CRRT (previous).

Table 1 Evolution of hemodynamic parameters (mean ± SD)

	Previous	Initial	30 min	60 min	90 min	120 min
HR p	125.2 (24.2)	124.8 (29.9) 0.1	128.8 (29) 0.07	129.4 (27.7) 0.03	129.1 (24.8) 0.03	131.8 (25.9) 0.01
MAP p	106.9 (21.6)	107.1 (19.5) 0.6	115.4 (19.3) 0.06	113.7 (18.6) 0.001	118.5 (28) 0.001	120.7 (20.5) 0.001
CVP p	7.6 (2)	6.1 (3) 0.04	6 (3.2) 0.01	5.6 (2.5) 0.001	5.6 (2.6) 0.001	5.1 (2.6) 0.001
MPAP p	17.1 (5.7)	19.6 (8.8) 0.03	21.1 (8.8)0.001	21.6 (6.4) 0.001	21.7 (6.5) 0.001	20 (6.3) 0.001
PAOP p	7.9 (3.7)	7.6 (4.7) 0.5	5.8 (4.4) 0.01	5.6 (4.3) 0.001	6.4 (5) 0.003	6 (4.2) 0.001
CFI p	9.1 (1.5)	9.2 (2.4) 0.3	8 (2.2) 0.2	9.3 (2) 0.7	9.3 (2.2) 0.7	9.3 (2.5) 0.8
SVI p	37.7 (7.7)	35 (8.5) 0.003	34.3 (6.7) 0.001	33.8 (7.9) 0.001	34.2 (6.7) 0.003	32.1 (8.5) 0.003
VVS p	16.6 (6.3)	18.9 (5.5)0.7	18.3 (5.9) 0.7	16.7 (4.9) 0.3	16.5 (6) 0.7	19.9 (6) 0.8
GEDV p	206.5 (49.1)	196.2 (46) 0.02	201.2 (41.6) 0.2	196.1 (50.5) 0.005	199.1 (51.3) 0.05	184.1 (47.2) 0.002
EVLWI p	15.9 (3.9)	15.1 (4.2) 0.08	15.4 (3.5) 0.1	16.6 (6.2) 0.7	17.2 (5.6) 0.8	16.3 (5.6) 0.2
Temperature p	34.9 (1)	34.5 (1.2) 0.01	34.3 (1.2) 0.003	34.1 (1.4) 0.001	33.9 (1.5) 0.001	33.8 (1.5) 0.001

HR, heart rate (bpm); MAP, mean arterial pressure (mmHg); CVP, central venous pressure (mmHg); MPAP, mean pulmonary arterial pressure (mmHg); PAOP, pulmonary arterial occlusion pressure (mmHg); CFI, cardiac function index (L/min); SVI, stroke volume index (mL/m²); SVV, systolic volume variation (%); GEDV, global end-diastolic volume (mL); EVLWI, extravascular lung water (mL/m²). Temperature (C°).

*p: statistical comparison with values before CRRT (previous).

Figure 2. Evolution of systemic vascular resistance index (SVRI) (dyn × cm⁵/m²) (p < 0.05) (analysis of variance [ANOVA] for repeated measures). *p: statistical comparison with values before CRRT (previous).

Figure 3. Evolution of left ventricular contractility (Dp/dtmax) (mmHg/s) (p < 0.05) (analysis of variance [ANOVA] for repeated measures). *p: statistical comparison with values before CRRT (previous).

Figure 4. Evolution of cardiac output (CO) (L/min) (p < 0.05) (analysis of variance [ANOVA] for repeated measures). *p: statistical comparison with values before CRRT (previous).

There were no significant differences in any hemodynamic variables when the three blood flows (5, 15, and 30 mL/min) were compared. We found no differences in hemodynamic variables between the two CRRT pumps or between hemofiltration and hemodiafiltration. The negative fluid balance obtained with CRRT during the 2 h was 11 ± 14 mL. There were no significant changes in blood gases, hematocrit, and electrolyte values during the experiment.

DISCUSSION

In children, hypotension is normal at the beginning of venovenous CRRT. Hypotension is more frequently observed in neonates, infants, and patients with previous hemodynamic alteration. The initial hypovolemia is in relation to the priming volume of the circuit and filter size. In pediatric patients, and moreover in infants and neonates, it is important to use circuits and filters with low priming volume to reduce initial hypovolemia. However, our study shows that venovenous CRRT in a healthy pediatric animal model induces an initial hypovolemia due to blood replenishment of extracorporeal circuit, despite circuit and filter low volume. Hypovolemia was associated with a significant decrease of volemia-related parameters (GEDVI, ITBVI), and cardiac filling pressure parameters (CVP, PAOP). Vasoconstriction and increased ventricular contractility were observed as cardiovascular response to maintain blood pressure. CO diminished secondary to SVRI increase, although CFI (the ratio of CO to the GEDVI) did not change. These hemodynamic changes maintained during the 2 h of CRRT probably due the increase of venous capacitance with secondary relative hypovolemia. Controlled ultrafiltration is always used in clinical practice in CRRT, and zero-balance CRRT would be more appropriate to study the effect of connection of the CRRT circuit. Although we did not perform controlled ultrafiltration in our experiment, the negative balance obtained during the 2 h was low. Our results in this pediatric animal model are in accordance with those published in adults.[4–7] Adult patients with venovenous CRRT show better hemodynamic tolerance than with intermittent hemodialysis.[4–7] Venovenous CRRT induces vasoconstriction and increased plasma norepinephrine levels, whereas hemodialysis induces vasodilatation and hypotension.[11]

There are several methods to measure CO in clinical practice. Pulmonary arterial thermodilution by Swan-Ganz catheter is the most used method to measure CO in adult critically ill patients. However, in children, the insertion of Swan-Ganz catheter is more difficult than in adults, and the risk of complications is higher. Femoral arterial thermodilution is a new invasive method to measure CO by means of a femoral artery and a central venous line. This method has showed good agreement with pulmonary arterial thermodilution technique in several studies in adults,[12],[13] children,[14] and baby animals.[15] Moreover, analysis of the femoral arterial thermodilution curve provides information of cardiac filling volumes, SVI, SVV, GEDVI, ITBVI, and EVLWI. We did not find studies that analyze the changes associated with venovenous CRRT in volemia-related parameters measured by femoral arterial thermodilution.

Several studies in adults and children have shown that GEDVI, ITBVI, and SVI measured by PiCCO are better preload indicators than CVP or PAOP.[16],[17] CVP and PAOP measure the repercussion of volemia on pressure. Pressure does not accurately correlate with volemia in patients with ventricular dysfunction, distensibility ventricular alterations, or high intrathoracic pressure (mechanical ventilation with high PEEP). In our study with healthy animals, both volemia and related pressure parameters decreased at the beginning of venovenous CRRT. The SVV can be a predictor of fluid responsiveness.[18] In our study, this parameter initially increased, probably secondary to hypovolemia, but the changes were not significant. CFI provides an accurate estimation of left ventricular systolic function.[19] This parameter did not change in our animals, although there was an increase in ventricular contractility showing that the most important hemodynamic adaptation change was an increase in vascular resistance.

We found that CRRT induces a slightly increase in PAOP. An experimental study showed that CRRT did not alter pulmonary hemodynamic parameters in healthy animals, but increased pulmonary vascular resistance in the presence of sepsis.[20]

CRRT improves oxygenation, lungs mechanics, and pulmonary edema in animals with lung inflammation.[21] In our healthy animal model, EVLWI did not change during the study, suggesting that negative fluid balance was intravascular.

Venovenous CRRT induces a decrease in core temperature with subsequent increase of systemic vascular resistance.[4],[5],[22–24] The cooling is caused by blood flowing into the extracorporeal circuit, and the temperature of replacement and dialysate fluid.[5] Cooling is a major problem in children and when high volume circuit and fast blood flows are used.[22],[23] However, we did not found significant differences in temperature changes between the three different blood flows. In our study, the decrease in rectal temperature fall could be partially responsible of the increase in vascular resistance and blood pressure. Although hypothermia usually induces bradycardia, in our animals, this response could be compensated by the increase in heart rate induced by hypovolemia. The fall in core temperature observed in our study was moderate because the duration of CRRT was only 2 h. Nevertheless, in long-lasting CRRT, temperature control is very important to prevent severe hypothermia that can induce low cardiac output.

High blood flows can be associated with more hemodynamic alterations because of higher blood volume percentage flowing into the extracorporeal circuit. We did not found significant differences in hemodynamic parameters between the three different blood flows, perhaps because they are in the low range used in the clinical practice.

However, the more important factors associated with hemodynamic alterations due to CRRT are the circuit and filter volume, the weight of the patient, previous hemodynamic status, and negative fluid balance.

Our study analyzed the hemodynamic response to CRRT in healthy animals. Nevertheless, most critically ill children with CRRT suffer pathologies that alter the hemodynamic response to hypovolemia and hypothermia. These patients have frequently hypervolemia and low blood pressure despite treatment with cardiovascular drugs. In a study with septic shock adult's patients,[4] CRRT induced similar hemodynamic changes to those observed in our study. Although mean blood pressure increased, 45% of patients treated with CRRT presented with hypotensive episodes.[4] Reeves did not find significant hemodynamic changes with arteriovenous and venovenous hemofiltration in healthy pediatric-size animals, but the changes were more important in animals with sepsis.[25]

In conclusion, we found that CRRT induces hypovolemia in the moment of circuit connection that leads to an increase in SVRI and left ventricular contractility, and is accompanied by a decrease in CO. Controlled ultrafiltration should be used when performing CRRT in children to reduce iatrogenic hypovolemia. Critically ill children with CRRT need careful monitorization to prevent and treat hemodynamic alterations. Studies that analyze the hemodynamic alterations in critically ill children with CRRT are needed.

ACKNOWLEDGMENTS

This work was supported by a FIS grant: No. 00/0013-06. We thank Francisco José del Cañizo, Mercedes Adrados Plaza, and Angélica Biurrun González of the Experimental Medicine Department of Gregorio Marañón Hospital for their help in the experiments, as well as Jose María Bellón for the statistical analysis.