Continuous arteriovenous hemofiltration after cardiac operations in infants and children

Continuous arteriovenous hemofiltration after cardiac operations in infants and children Acute renal insufficiency after cardiopulmonary bypass can lead to a significant morbidity from fluid overload and electrolyte disturbance, impede pulmonary gas exchange, and postpone weaning from mechanical ventilation. The limitations placed on free water intake result in severe restriction of nutrition while diuretic therapy causes electrolyte imbalance. Artificial renal support either in the form of peritoneal dialysis or hemodialysis may be complicated by sepsis and hemodynamic instability. We reviewed our experience with the use of continuous arteriovenous hemofiltration, an extracorporeal technique for removal of solutes, toxins, and water in critically ill patients with cardiac failure complicated by acute renal insufficiency and hemodynamic instability after cardiopulmonary bypass. Ten infants and children with renal insufficiency caused by low cardiac output had continuous arteriovenous hemofiltration instituted for indications including sepsis, volume overload, oliguria for more than 24 hours nonresponsive to diuretic therapy, and the need for hyperaiimentation. All were supported by mechanical ventilation and receiving high-dose inotropic support. Arterial and venous vascular access was successfully obtained by cannulation of the femoral artery and vein in nine patients. Anticoagulation of the circuit was achieved with heparin infusion (6 to 20 ~g/kg/hr) and monitored by measurement of activated clotting time. The continuous arteriovenous hemofiltration circuit was replaced if there was clot formation, or at 3 days after placement. Dialysis solution (Dianeal) 1.5 % or 0.5% was infused as prefilter dilution. With the use of continuous arteriovenous hemofiltration, 20 to 100 m/hr of ultrafiltrate was removed, which aDowed correction of hypervolemia, and caloric intake increased from 13.5 kcal/kg/day to 79.5 kcal/kg/day, Continuous arteriovenous hemofiltration was maintained between 5 hours and 8 days and was weD tolerated in all patients. Serum urea and creatinine levels declined during continuous arteriovenous hemofiltration. We conclude that continuous arteriovenous hemofiltration is a safe and effective method for fluid and electrolyte homeostasis and that it thus allows hyperaiimentation in infants and children after cardiac operations. (J THORAC CARDIOVASC SURG 1992;104:1225-30)

Gideon Paret, MD, Amram J. Cohen, MD, Desmond J. Bohn, MB, FRCPC, Helen Edwards, RN, Robert Taylor, MB, FFARCSI, Denis Geary, MB, FRCPC, and William G. Williams, MD, FRCSC, Toronto, Ontario, Canada

Acute renal insufficiency in the postoperative period is one of the most common complications of cardiopulmonary bypass in infants and children. I, 2 Although in most instances the situation resolves spontaneously within 48 From the Departments of Critical Care and Pediatrics (Nephrology) and the Division of Cardiovascular Surgery, The Hospital for Sick Children, Toronto, Ontario, Canada. Read at the Sixteenth Annual Meeting of The Western Thoracic Surgical Association, Coronado, Calif., June 20-23, 1990. Address for reprints: Desmond Bohn, MB, FRCPC, The Pediatric Intensive Care Unit, The Hospital for Sick Children, 555 University Ave., Toronto, Ontario M5G lX8, Canada.

12/6/40514

hours, where it does not, the resulting hypervolemia and electrolyte disturbance can lead to significant morbidity and occasional mortality. The need for artificial renal support (hemodialysis or peritoneal dialysis) in patients after heart operations has been associated with a mortality rate of 50% to 70%.3-6 In addition, excess total body and lung water can impede pulmonary gas exchange and postpone weaning from mechanical ventilation, and the limitations placed on free water intake result in severe restriction of nutritional support. Fluid overload and electrolyte disturbance that cannot be managed with diuretic therapy can be treated with artificial renal support in the form of either peritoneal dialysis or hemodi1225

The Journal of Thoracic and Cardiovascular Surgery

1 2 2 6 Paret et al.

ARTERIAL

LINE

VENOUS LINE FILTER

HEPARIN

REPLACEMENT FLUID

ULTRA FILTRATE

Fig. 1. System diagram of CA VH. Arrows indicate direction of flow. Amount of fluid filtered is controlled by infusion pump on ultrafiltrate line.

Table I. Guidelines for cannula size and type offilter

to be used for CAVH Patient weight

Cannula size

Patient weight

Filter

<5 kg 5-15 kg >15 kg

5F 7F 8F

<5 kg 5-25 kg >25 kg

Gambro FH-22* Amicon Diafilter rot Amicon Diafilter 20t

*Gambro Hospal, Inc., Williamsburg,Va, tAmicon Division, W.R. Grace & Co., Beverly, Mass.

alysis, although neither is entirely satisfactory. In the case of peritoneal dialysis, it can be difficult to achieve a satisfactory negative balance, even with a surgically placed tenkoff catheter, because of either leakage around the insertion site or inadequate peritoneal perfusion in addition to the increased risk of infection. Hemodialysis can only remove fluid on an intermittent basis and not infrequently results in hemodynamic instability. In 1977 Kramer and associates? described the alterna-

tive technique of continuous arteriovenous hemofiltration (CAVH) for correcting fluid and electrolyte imbalance, a system that provides for continuous fluid removal by ultrafiltration with gentle hemodynamic and electrolyte modulations. Throughout the 1980s this technique and its modifications have been used with increasing frequency in intensive care units in both children and adults. 8- 12 We have reviewed our experience in applying CAVH in children with acute renal insufficiency and fluid overload after cardiac operation to evaluate the efficiency of fluid removal and solute clearance and our ability to maintain the patients in nutritional balance. Methods Our study included a retrospective review of all children in whom CAVH was used to manage acute renal insufficiency after heart operations during a 3-year period. Indications for treatment included (1) persistent anuria or oliguria ( < 1mljkgj hr) despite supportive and diuretic therapy, (2) evidence ofuremia, (3) diuretic-resistant hypervolemia, or (4) the necessity for intravenous nutrition in the presence of decreased renal function and positive fluid balance. Demographic data including age,

Volume 104 Number 5 November 1992

CAVH

1227

Table II. Complications, outcome, duration of treatment, andjluid balance in nine patients managed with CAVH Patient

Age (yr)

Cardiac diagnosis

Operation

9

PA,VSD,AS

2

1.8

TGA, OILV, PS, VSD

ClosureVSD, replacement homograft, aortic valvotomy Fontan procedure

3

0.1

VSD, coarctation

ClosureVSD

4

0.1

VSD, aortic atresia

5

4

PA,VSD

6

3

Tetralogy, AVSD

7

7

Tetralogy, AVSD

8

1.3

PS, VSD, pulmonary arterial sling

9

3

PA, VSD, single coronary artery

Aortic arch repair, closure VSD Conduit placement, VSD repair Repair AVSD, RVpulmonaryarterial conduit Two operations, repair, closureresidual VSD Two operations, pulmonary valvotomy, closureVSD, tracheal resection ClosureVSD, RV-pulmonary arterial conduit,RPA plasty

Coincident pathologic conditions

Outcome

Acute hepatic failure, single kidney

Survival

Raised LFTs, metabolic acidosis S. epidermidis sepsis, dysplastic kidneys S. epidermidis, Pseudomonas sepsis S. epidermidis sepsis

Death

Survival

S. epidermidis sepsis

Death

Cardiac arrest

Death

Singlekidney, tracheal stenosis

Death

Pseudomonas sepsis, portal

Death

Survival Death

hypertension, esophageal varices

PA, Pulmonary atresia; VSD, ventricular septal defect; AS, aortic stenosis; TGA, transposition of the great arteries; DILV, double-inlet left ventricle; PS, pulmonary stenosis; A VSD, atrioventricular septal defect; RPA, right pulmonary artery.

weight, and sex were recorded for each patient. The clinical condition of each patient including the need for mechanical ventilation, inotropic support, hemodynamic instability, and evidence of other system failure was recorded. Once CAVH was begun the success or failure of the treatment was noted. Success was defined as the ability to remove adequate ultrafiltrate to control the patient's fluid balance, electrolyte status, and uremia. Duration of treatment was recorded as was the number of filters used for each patient. Filters were changed when there was evidence of clot formation, or every 3 days. Mean values before and after CAVH were recorded for urea, creatinine, and caloric intake for each patient. The complications related to the procedure and the clinical outcome for each individual patient were recorded. Technical principles. Hemofiltration is an extracorporeal form of organ support whereby fluid is removed in the form of an ultrafiltrate of plasma through a highly permeable artificial membrane that allows free passage of fluid and electrolytes, but remains relatively impermeable to plasma proteins. Plasma ultrafiltrate loss is balanced by a combination of a prefiltration electrolyte solution and intravenous alimentation fluid. The amount of fluid filtered across the membrane is governed by the hydrostatic pressure (arterial minus venous) across the membrane and can be controlled by altering the pressure in the ultrafiltrate line. The net fluid balance is determined by the difference between the total fluid input (prefiltration fluid plus alimentation fluid) and the amount of ultrafiltrate. Blood coagulation on the filter is prevented by heparinization and maintenance of a high blood flow through the extracorporeal circuit. The typical circuit used for hemofiltration is depicted in Fig. 1. Vascular access was established in this series with polyethyl-

ene hemofiltration cannulas (Cook Incorporated, Bloomington, Ind.) via the femoral vessels. Percutaneous access with a Seldinger technique was used where possible, the alternative of a cut-down being used when this failed. The largest cannula sizes used according to body weight are given in Table I. The cannulas are short with minimal taper, which provides the least resistance. The cannulas are connected to hemofiltration tubing (Amicon Division/W.R. Grace & Co., Beverly, Mass.). The length of the arterial tubing is 50 cm and the venous tubing length is 65 em. Internal diameter of the tubing is 4.79 mrn. Two side ports are connected to the arterial side; one for a continuous infusion of heparin, the other for a prefiltration replacement fluid. Amicon Diafilter 10, Amicon Diafilter 20, and Gambro FH-22 filters (Gambro Hospal, Inc., Williamsburg, Va.) were used in this series (Table I). Ultrafiltrate removal is controlled by passing the tubing through a gravimetric pump that is set to restrict flow to a preselected rate. The blood is returned to the patient via the venous line. No side ports are used for the venous side of the system. Fluid and electrolyte control. All fluids added to and removed from the system are controlled by infusion pumps. Heparin is added to maintain an activated clotting time between 175 and 200 seconds (Hemochron, International Technidyne Corp., Edison, N.J.). Prefiltration replacement fluid is given as 0.5% or 1.5% dialysis solution (Dianeal, Baxter Healthcare Corp., Deerfield, Ill.), with or without added potassium depending on the patient's electrolyte status. All other basic electrolyte requirements are contained in the intravenous alimentation solution. The prefiltration replacement solution is started at a rate of 3 to 5 ml/rnin, The blood is then returned to the patient via the venous conduit. The net hourly negative fluid balance is determined by subtracting the sum of the fluid input (prefiltra-

The Jourrial of Thoracic and Cardiovascular Surgery

1 2 2 8 Paret et al.

40

•

0

UREA PRE UREA POST

30

~

~

..

20

.§.

W 0:

::>

10

A 5 0 0 , - - - - - - - - - - - - - - - -_ _-, •

400

o

CREATININE PRE CREATININE POST

~

"0

,E

300

w

Z

Z

>= ~

200

0:

U

100

B Fig. 2. Change in urea (A) and creatinine(B) in nine patients before and after successful establishmentof CAVH.

tion replacementfluidplusintravenous infusions) from the predetermined amount of ultrafiltrate. Monitoring. The electrolytestatus and hematocritvalueare monitored every 4 hours; prothrombin time, partial thrombin time, and platelets every 12 hours. Urea is measured every 12 hoursand creatininedaily.Thoughthe set up and primingof the CAVH circuit is done by dialysis nurses, intensive care unit nursingstaffoperate the systemwithout any additionalspecialized personnel.

Results From January 1988 to December 1989, 10 children who underwent corrective cardiac operations had CAVH instituted. There were six male and four female patients. Mean age was 2.94 years (range 1 month to 9 years) and mean weight was 11.43 kg (range 2.9 to 20.9 kg). The cardiovascular diagnosis, other medical conditions, and details of the surgical procedures of the nine patients who were successfully placed on CAVH are shown in Table II. Three of the children had preexisting renal abnormalities; two with single kidneys (one of whom also had portal hypertension and varices) and one with

renal dysplasia. One further child had a long-segment tracheal stenosis. In six of the patients total bypass time was more than 2 hours with a maximum of 5 hours. Three children in this series required a second surgical procedure to repair residual anatomic defects (closure of a ventricular septal defect and placement of a pulmonary homograft with right ventricular muscle bundle resection). All patients had continuing requirements for mechanical ventilation and inotropic support. Three had an open sternum and five had sepsis with positive bacterial identification on blood culture (three Staphylococcus epidermidis and two Pseudomonas). Acute hepatic failure also developed in one child in the postoperative period. In one patient renal insufficiency was managed with peritoneal dialysis before CAVH was instituted. CAVH was started an average of 10 days (2 to 30 days) after operation. In only two patients was the procedure started later than 10 days postoperatively, and in five patients it was started within 5 days of the operation. The institution of hemofiltration was successful in nine children. In one patient, a l-month-old child weighing 2.9 kg, CAVH was not technically successful because of difficulty in gaining adequate vascular access. One further child died within 2 hours of the start of treatment. Treatment duration averaged 57 hours, and the longest treatment course was 172 hours. There were decreases in urea (mean 20.2 ± 8.6 to 12.9 ± 7 mmol/L) and creatinine (mean 145 ± 116 to 113 ± 83 JLmol/L) during treatment in the majority of patients (Fig. 2, A and B), although the primary goal of CAVH in most patients was control of volume status rather than the treatment of azotemia. Negative fluid balance per patient averaged 36 ml/hr, The maximum negative fluid balance achieved in anyone patient was 100 ml/hr. When the filters were functioning the amount of filtrate removed did not limit therapy in any patients. Caloric intake went from 13.5 ± 5.2 kcal/kg/day before CAVH to 79.5 ± 44 keel/kg/ day during CAVH (p = 0.003, Fig. 3). This was provided as intravenous alimentation fluid in the form of 10% to 30% dextrose with protein together with either 10% or 20% lipids. As our experience grew, our ability to give a large amount of calories with this technique improved. The most recent five patients all received more than 85 kcal/kg/day. Two patients had sufficient bleeding from the cannula site to require transfusion. In no patient did the hemorrhage require discontinuation of the technique or surgical correction. There was no evidence of the development of vascular cannula sepsis during CAVH. In one patient ischemia developed in the leg on the side of the arterial line. The patient was moribund with a low-flowstate and died within 2 hours of the event. In no patient could

Volume 104 Number 5 November 1992

hemodynamic instability be attributed to the use of CAVH. A total of 16 filters clotted during our experience. Each child required an average of 2.6 filters during treatment, but we could not relate clotting to the type of filter used. The amount of blood lost in the extracorporeal circuit with each clotted filter depended on the type of filter used. Three patients survived to leave the intensive care unit. Two had return of normal renal function and one required long-term hemodialysis.

CAVH

1229

200

• 0

CALORIES PRE CALORIES POST

150 >c

-e

0-

'" ;;; ....

100

'" 0

" 50

Discussion Fluid overload in children after cardiac operations is frequently seen in association with acute renal insufficiency. Attempts to resolve the problem by fluid restriction and pharmacologic diuresis are often not only unsuccessful but can also lead to a negative nitrogen balance together with large electrolyte losses. The result is a cycle of increasing edema, nutritional failure, and failure to wean from mechanical ventilation. The dilemma of providing an adequate caloric intake in the face of fluid overload makes parenteral nutrition impractical without some form of extracorporeal fluid removal. Gailiunas and associates'! have demonstrated that early aggressive hemodialysis reduces mortality in adult patients with acute renal failure after cardiopulmonary bypass. However, conventional hemodialysis in this setting is frequently complicated by cardiovascular instability caused by fluid shifts, which not only hinders attempts to attain negative fluid balance, but may also contribute to further ischemic renal injury. Severe electrolyte shifts have also been described.lv'" These problems are magnified in the hemodynamically labile child after cardiac operations. While peritoneal dialysis causes relatively less hemodynamic disturbance, it has the disadvantage of being relatively inefficient with low-solute clearances. Three patients in our series were switched from peritoneal dialysis to CAVH because of the inability to achieve a negative fluid balance. Peritoneal dialysis can also cause respiratory compromise from mechanical restriction of diaphragmatic excursion and has the additional potential complications of hyperglycemia, dialysate leak, and protein loss.!" Escape of dialysate fluid into the pleural cavity has been observed in children after cardiac operations. CAVH, on the other hand, permits accurate control of fluid balance in oliguric pediatric patients, so that intravascular space can be created for early aggressive nutrition and other fluid therapy. This is achieved by continuous ultrafiltration, which minimizes disequilibrium syndromes and hypotension associated with intermittent hemodialysis. 17 Several studies have demonstrated the importance of

Fig. 3. Caloric intake beforeand after institutionof CAVH in nine patients.

providing adequate nutritional support in severely ill patients. Abel and associates'" demonstrated improved survival with parenteral nutrition in acutely ill patients with renal failure. Kramer and co-workers'Y" have shown improved survival in acutely ill patients with renal failure with the use of CAVH and parenteral nutrition. All children in our series had complex congenital heart disease and were critically ill in a low cardiac output state because of either poor myocardial function after bypass or residual anatomic defects, complicated in several instances by septicemia. The importance of recognizing preexisting congenital renal abnormalities as a cause of postoperative renal dysfunction is demonstrated by the fact that three of the patients in this series had such findings. We have been able to demonstrate that in this population of severely ill children, fluid imbalance was corrected at the same time that there was a significant increase in caloric intake. Although the primary purpose of CAVH is fluid removal rather than dialysis, it is worth noting that the serum creatinine level was significantly reduced in 50% of the patients. The efficiency of dialysis can be improved in this system by incorporating a dialysis technique in which the membrane itself is dialyzed. No disequilibrium syndromes were noted during CAVH and no episodes of hypotension could be related to the use of CAVH. Complications were all related to catheter placement. CAVH is a safe and simple method for the treatment of renal insufficiency after cardiac operations in children. It is more efficient at removing fluid than peritoneal dialysis with a reduced risk of septic complications. Being a continuous rather than intermittent method of removing fluid, it causes less hemodynamic instability than hemodialysis. The technique allows for aggressive hyperalimentation in children with multiple organ failure without

I 2 3 0 Paret et al.

the complication of fluid overload. The only limitations to its use are an inadequate mean arterial pressure ( <45 mm Hg), which results in too Iowa perfusing pressure across the membrane, and inability to gain vascular access. Although the mortality was high in these patients with this early limited experience, this reflects more the severity of the underlying disease than any fault with the technique. Weare hopeful that the ability to safely control volume status in these patients earlier in the postoperative period and the ability to provide adequate nutritional support in this setting wi11lead to improved survival in the future. REFERENCES 1. Abel RM, Buckley MJ, Austen WG, Barnett GO, Beck CH, Fischer JE. Etiology, incidence, and prognosisof renal failure following cardiac operations. J THORAC CARDIOVASC SURG 1974;71:323-33. 2. Chesney RW, Kaplan BS, Freedom RM, Haller JA, Drummond KN. Acute renal failure: an important complication of cardiac surgery in infants. J Pediatr 1975;87:381-

8. 3. Bhat JG, Gluck ML, Cowenstein J, Baldwin DS. Renal failure after open heart surgery. Ann Intern Med 1976; 84:677-82. 4. Yeboah ED, Petrie A, Pead JL. Acute renal failure and open heart surgery. BMJ 1972;1:415-8. 5. Rigden SPA, Barrett TM, Dillon MJ, De Laval M, Stark J. Acute renal failure complicating cardiopulmonary bypass surgery. Arch Dis Child 1982;57:425-30. 6. Hodson EM, Kjellstrand CM, Maurer SM. Acute renal failure in infants and children: outcome of 53 patients requiring hemodialysis treatment. J Pediatr 1978;93:756-

61. 7. Kramer P, Wiggen W, Rieger J, Mathaei D, Scheler F. Arteriovenous hemofiltration: a new and simple method for treatment of overhydrated patients resistant to diuretics. Klin Wochenschr 1977;55:1121-2.

The Journal of Thoracic and Cardiovascular Surgery

8. Leiberman KY. Continuous arteriovenous hemofiltration in children. Pediatr NephroI1987;1:330-7. 9. Leiberman KY, Nardi L, Bosch JP. Treatment of acute renal failure in an infant using continuous arteriovenous hemofiltration. J Pediatr 1985;106:646-9. 10. Leone MR, Jenkins RD, Golper TA, Alexander SR. Early experience with continuous arteriovenous hemofiltration in critically ill pediatric patients. Crit Care Med 1986; 14:1058-63. II. Stevens PE, Davies SP, Brown EA, Riley B, Gower PE, Kox W. Continuous arteriovenous hemodialysisin critically ill patients. Lancet 1988;2:150-2. 12. Lopez-Herce J, Dorao MA, Delgado L, Espinosa F, Ruza F, Martinez Me. Continuous arteriovenous hemofiltration in children. Intensive Care Med 1989;15:224-7. 13. Gailiunas P, Chawla R, Lazarus JM, Chan L, Sanders J, Menill JP. Acute renal failure followingcardiac operations. J THORAC CARDIOVASC SURG 1980;79:241-3. 14. Tzamalouhas AH, Garella S, Chazan A. Peritoneal dialysis for acute renal failure after major abdominal surgery. Arch Surg 1973;106:639-43. 15. Myers BD, Morin SM. Hemodynamically mediated acute renal failure. N Engl J Med 1986;314:97-105. 16. Paradiso C. Hemofiltration: an alternative to dialysis. Heart Lung 1989;18:282-90. 17. Schurek JH, Biela D. Continuous arteriovenous hemofiltration: improvement in handling of fluid balance and heparinization. Blood Purif 1983;I:89-196. 18. Abel RM, Beck CH, Abbott WM, Ryan JA, Barnett GO, Fischer JE. Improved survival in renal failure after treatment with intravenous essential L-amino acids and glucose. N Engl J Med 1973;288:695-9. 19. Kramer P, Seezers A, De Vivie R. Therapeutic potential of hemofiltration. Clin NephroI1979;1l:145-9. 20. Kramer P, Bohler J, Kehr A, et al. Intensive care potential of continuous arteriovenous hemofiltration. Trans Am Soc Artif Intern Organs 1982;28:28-32.