Kramer, Robert E.*; Sokol, Ronald J.*†; Yerushalmi, Baruch*; Liu, Edwin*; MacKenzie, Todd†; Hoffenberg, Edward J.*; Narkewicz, Michael R.*
*Pediatric Liver Center and Liver Transplantation Program, Section of Pediatric Gastroenterology, Hepatology and Nutrition, Department of Pediatrics, and the †Pediatric General Clinical Research Center, The Children’s Hospital and University of Colorado Health Sciences Center, Denver, Colorado, U.S.A.
Received September 11, 2000;
revised March 22, 2001 and May 3, 2001; accepted May 17, 2001.
Presented at the 2000 World Congress of Pediatric Gastroenterology, Hepatology and Nutrition, August 5–9, 2000, Boston, MA, and published in abstract form (JPGN 2000;31(2):S113).
Address correspondence and reprint requests to Dr. Ronald J. Sokol, The Children’s Hospital, Box B-290, 1056 East 19th Avenue, Denver, CO 80218, U.S.A. (e-mail: email@example.com).
Ascites is a frequently encountered complication of cirrhosis in both children and adults. The management of ascites includes salt and fluid restriction, treatment with diuretics, surgical shunt placements, transjugular intrahepatic portosystemic shunting, and liver transplantation. In adults, large-volume paracentesis (LVP) is an additional safe and effective technique for the management of tense, unresponsive abdominal ascites (1). In fact, pulmonary function testing shows marked improvement after LVP in patients with pulmonary compromise resulting from tense ascites (2,3).
The first report of the use of paracentesis in children was by Denzer in 1920 (4), who advocated its use for diagnosis in pediatric peritonitis. Subsequently, paracentesis has been advocated for the diagnosis of urinary (5), cardiac (6), traumatic (7), and chylous ascites (8) in infants and children. However, the use of LVP as a treatment for ascites in children has not been reported. We have used this technique for the past 3 years in children with tense ascites poorly responsive to other treatments. The objectives of this initial report were to review the safety and efficacy of LVP in the management of ascites and to compare the speed and efficacy of LVP using two types of catheters.
MATERIALS AND METHODS
This study was a retrospective review of all LVPs performed at The Children’s Hospital by the Department of Pediatric Gastroenterology, Hepatology and Nutrition between January 1, 1997 and June 16, 2000. For the purpose of this study, LVP was defined as removal of 50 ml or more of ascitic fluid per kilogram of dry body weight. There were a total of 21 LVP sessions performed in seven children that satisfied the above criteria. All seven patients had tense abdominal ascites with either abdominal discomfort or respiratory compromise. All seven had been treated with combination diuretics ranging from 0.5 to 2.0 mg/kg per day for furosemide and 0.3 to 2.25 mg/kg per day for spironolactone. Six of the seven were treated with a standard sodium restricted diet limited to 1 to 2 mEq/kg per day for the infants and children and 1 to 2 g/day in the adolescents.
Large-Volume Paracentesis Technique
Informed consent was obtained from patients or families before each LVP procedure. Patients fasted for at least 2 to 4 hours while receiving maintenance intravenous fluids before the procedure. The LVP procedure was performed with the patient in the supine position and usually under conscious sedation with intravenous midazolam, meperidine, or both. Baseline complete blood count (n = 19) and prothrombin time (n = 12) were obtained before the procedure. Blood products were administered on an individual basis; platelets were typically infused if platelet count was less than 50,000, and fresh frozen plasma was given if prothrombin time was prolonged more than 5 seconds above normal. All patients underwent abdominal ultrasonography before the initial LVP to locate a safe site for paracentesis. Local analgesia was achieved by administration of topical 2.5% lidocaine and 2.5% prilocaine cream (EMLA cream; Astra Pharmaceuticals, Wayne, PA) 30 to 60 minutes before the procedure or by infiltrating the site with 1% lidocaine immediately before catheter insertion. Either a 16-gauge (n = 6) or 18-gauge (n = 3) Teflon intravenous catheter [IC] (Quik-Cath; Baxter Healthcare Corporation, Deerfield, IL) or a 15-gauge (n = 12) paracentesis needle [PN], (Caldwell needle; Ballard Medical Products, Draper, UT) was used, depending on physician preference. The PN consists of an 8.25-cm stainless steel canula with a blunt tip, two fenestrations along each side, a removable beveled stylet that sits inside the canula, and a flared lip at the top, allowing it to be connected to a standard three-way stopcock device. Insertion was performed using the “Z technique” of repositioning the needle before entry into the peritoneal cavity whenever possible to minimize leakage after removal of the catheter. After return of ascitic fluid, the stylet was removed and the catheter left in place and attached to a three-way stopcock. Samples were collected for laboratory testing, and then the stopcock was connected to a sterile collection system for drainage by gravity. For the IC technique, the stopcock assembly was covered with a paper cup and taped to the abdomen during drainage to prevent dislodging by the child. This was not usually necessary for LVPs performed using the PN because of the much shorter collection time. Vital signs, oxygen saturation, and urine output were recorded every 15 minutes during the procedure. Infusions of intravenous albumin were also administered on an individual basis to provide hemodynamic stability and replacement for removed ascitic protein (albumin); 0.5 to 1.0 g of 5% albumin per kilogram of dry weight were infused over 1 to 2 hours beginning at the time of catheter insertion. The choice of 5% over 25% albumin was based on efficacy of intravascular volume expansion. The paracentesis catheters were removed when drainage stopped or slowed considerably, and a pressure dressing was applied.
Data obtained from patient records for this study included the volume of ascites removed, the duration of drainage, the use of ultrasound guidance, needle type, complications, use of intravenous albumin, ascitic fluid analysis results, and coagulation status of the patients. Ascitic fluid volume removed was expressed as total volume per kilogram dry body weight, volume per hour, and volume per kilogram of dry body weight per hour.
Data were analyzed using standard Sigmastat statistical software (SPSS Inc., San Rafael, CA) and are presented as mean ± standard deviation. Comparisons between the PN, 16-gauge IC, and 18-gauge IC were performed using one-way analysis of variance, with Kruskal-Wallis analysis of variance where appropriate, and post hoc Tukey two-sample t test for pairwise comparison. Statistical comparison of the PN versus the combined IC (16 gauge plus 18 gauge) was also performed using the Student’s t test. Statistical significance was considered a P value < 0.05.
Seven patients ages 7 months to 19 years (mean age, 7.8 years), underwent a total of 21 LVP procedures at our institution during the 3-year study period (Table 1). The first patient underwent three LVPs over a 7-month period for ascites caused by chronic liver allograft rejection and cirrhosis. The second patient underwent one therapeutic LVP for ascites caused by hepatic failure that developed as a complication of Epstein-Barr virus-associated hemophagocytic syndrome. The third patient underwent three LVP procedures over a 2-month period for ascites secondary to congenital hepatic fibrosis, after undergoing a renal transplant. His recurrent ascites resolved after placement of a spleno-renal shunt. The fourth patient underwent eight LVPs over a 2-month period for chylous ascites occurring after surgery for a retroperitoneal Kaposiform hemangioendothelioma. A peritoneovenous shunt was eventually placed, however, the postoperative course was complicated by cardiac and respiratory failure and the shunt was removed subsequently. She is stable on chemotherapy. The fifth patient underwent three LVPs over a 12-month period for ascites caused by Budd-Chiari syndrome. Ascites resolved after he received an orthotopic liver transplant. The sixth patient underwent one LVP for ascites that developed secondary to veno-occlusive disease after stem cell rescue for treatment of medulloblastoma. After initial LVP, veno-occlusive disease improved and the ascites were controlled with diuretic therapy alone. The seventh patient underwent two LVPs for ascites that developed after orthotopic liver transplantation for carbamyl phosphate synthetase deficiency. She was subsequently found to have a necrotic edge of the reduced-size liver allograft; ascites resolved after resection of this necrotic liver tissue. Aggregate data for all patients is shown in Table 2.
In three LVP sessions (14%), the absolute neutrophil count of the ascitic fluid was found to be more than 250/mm 3 , which was considered diagnostic for spontaneous bacterial peritonitis despite the fact that all cultures remained negative. All three were treated empirically with intravenous third-generation cephalosporin therapy. No bacteremia or sepsis was identified as a complication of LVP. Albumin concentration of ascitic fluid ranged from 1.0 to 3.0 g/dL (mean, 1.79 ± 0.71 g/dL) and protein concentration ranged from 743 to 4600 mg/dL (mean, 2,225 ± 1,318 mg/dL). Ascitic amylase levels ranged from 4 to 220 IU/L (mean, 37.8 ± 57.2 IU/L), and triglyceride levels ranged from 10 to 972 mg/dL (mean, 215.8 ± 255.8 mg/dL). Results of ascitic cell counts are given in Table 1.
Comparison of catheters used for LVPs by one-way analysis of variance (Table 3) showed that there were significant differences between the three groups in flow rate (milliliters per hour;P = 0.001), weight-adjusted flow rate (milliliters per kilogram per hour;P < 0.001), and duration of drainage (hours;P = 0.004). Drainage volume per kilogram of body weight (milliliters per kilogram) was also greater with the PN than with either of the IC catheters, however, results did not achieve statistical significance (P = 0.2). Subsequent pairwise analysis of the groups showed that for milliliters per hour, milliliters per kilogram per hour, and hours there were significant differences between the PN and 16-gauge groups (P = 0.004, P < 0.001, and P < 0.001) but not between the PN and the 18-gauge groups (P = 0.17, P = 0.08, and P = 0.13), although this was expected because of the small number of LVPs in the 18-gauge group (n = 3). Therefore, a repeat comparison was performed using the Student’s t test after combining the 16-gauge and 18-gauge LVPs into a single IC group. It showed the PN to have significantly faster flow rate (milliliters per hour;P = 0.001), faster weight-adjusted flow rate (milliliters per kilogram per hour;P < 0.001), and shorter duration of drainage (hours;P < 0.001) than the IC group. Both triglyceride and protein concentrations of the ascitic fluid were not significantly different among the catheter groups (P = 0.16 and P = 0.45, respectively) and therefore were not believed to account for the differences observed in flow rates. All procedures were well tolerated, with the only complication being a mild decrease in urinary output in one patient, who responded well to volume expansion. There was no evidence of significant hemorrhage or leakage complicating any of the procedures.
Although LVP is frequently used in adults for the management of recurrent ascites resistant to medical therapy, its role in pediatric patients has not been defined. Our experience demonstrates that a large volume of ascitic fluid can be removed quickly, even from small infants, with minimal complications. The decreased urine output that was observed in one of our patients during LVP drainage was mild and responded to additional administration of intravenous albumin. Gines et al. (9) have performed controlled trials in adults comparing LVP with and without the use of albumin and found that intravenous albumin replacement reduces renal and electrolyte complications. Although alternative, less expensive volume expanders, such as Dextran 70, have been found to be equally efficacious as albumin in preventing these complications after LVP, use of Dextran 70 after LVP still resulted in increased renin and aldosterone activity (10). Use of alternative volume expanders in children has not been studied. Therefore, we chose to use albumin for both volume expansion and to avoid imposing any additional nutritional stress caused by removal of ascitic protein. Although nitrogen balance after serial LVP was not studied, in our experience, removal of ascites after LVP resulted in improved appetite and oral intake. We recognize that repeated removal of a large volume of ascites has the potential to cause protein depletion, and therefore recommend adequate oral protein intake and albumin infusions to prevent further protein losses. We initially chose to use 5% albumin to provide volume and albumin replacement concurrently. Inasmuch as we did not observe significant hemodynamic instability after LVP, we are now administering 25% albumin, as is used with LVP in adults (1).
Using the Z technique whenever possible, we did not observe any episodes of leakage from the LVP site, despite repeated procedures in the same area. The Z technique was more difficult to perform in the infants because of the relatively thin abdominal wall. Intraperitoneal hemorrhage, one of the potential complications, was not encountered in our series, despite the fact that coagulopathy and thrombocytopenia were present in a number of our patients. Based on ultrasound guidance, many of the LVPs were performed in either the right or left lower quadrants, where a greater potential risk of bleeding exists because of the vasculature of the rectus abdominus muscles. We chose to infuse fresh-frozen plasma before LVP if the PT was more than 5 seconds above the upper limit of normal for age, and platelets were given to maintain a platelet count of at least 50,000/mm 3 before and for 1 to 2 days after the procedure. In adults, however, recent studies show that the severity of thrombocytopenia or coagulopathy did not increase the risk of hemorrhage in LVP (11). Thus, the correction of coagulopathy in children undergoing LVP may not be required. The suspected mechanism of intraperitoneal hemorrhage after LVP is a rapid increase in the pressure gradient across the wall of mesenteric varices resulting from the sudden drop in intraperitoneal pressure with drainage, leading to rupture of the vessels. Mortality from hemorrhage of mesenteric varices is reported to be as high as 70% in adults, occurring anywhere between 3 hours and 4 days after LVP (12). In our patient with hemangioendothelioma and our patient with Budd-Chiari syndrome, the high number of red blood cells found in the ascitic fluid was believed to be caused by their primary illnesses. Neither had a significant drop in hematocrit after LVP.
The necessity of a preprocedure abdominal ultrasound has not been established. Although not all of our LVPs were under ultrasound guidance, at least one ultrasound examination was performed to determine an appropriate site before the first LVP was performed on all patients. For patients with a relatively short interval between LVPs, ultrasonography was not always repeated.
The optimal catheter system to be used for LVP in children had not been addressed before this report. Success using a peritoneal dialysis catheter system (13), in addition to the IC, has been reported in adults. Newer paracentesis needles have been developed to hasten flow rates of the drainage of ascites in a safe manner. In a study of adults with ascites comparing the PN with the IC, both connected to negative pressure, the PN had significantly faster flow rates, decreased need for subsequent paracenteses, and fewer premature terminations resulting from poor fluid return (14). Our experience with the PN draining to gravity was similar in that rate and duration of drainage were significantly improved compared with IC, without compromising safety. The fenestrations and rigidity of the PN allow better flow and therefore required less repositioning as drainage progressed. Although the faster rate of drainage with the PN may raise concerns about hemodynamic instability, this was not observed. Advantages of the shorter drainage time may be a reduction in the risk of infection and the ability of the physician to remain at the bedside throughout the procedure. The 3.75-inch (9.53-cm) length of the PC could not be inserted completely into the abdomen of the smaller patients, although it still worked well with only partial insertion as long as all of the fenestrated holes were below the surface of the skin. Although this was a retrospective study with a small number of patients that carries the risk of physician bias in catheter selection, the dramatic difference in flow rates observed between the PN and the IC argues against undue influence on the results.
In summary, LVP was both safe and effective in our initial experience in pediatric patients. Use of a fenestrated, stainless-steel paracentesis needle offered significant advantages over intravascular catheters in increasing the flow rate of drainage and shortening the duration of the procedure, as it does in draining ascites in adults. Using our technique, the only complication encountered was decreased urine output, which was readily amenable to volume expansion. Although prospective studies in adults with ascites have addressed the issues of volume expansion, rapidity of drainage, coagulopathy correction, and types of paracentesis needles to be used in LVP, these issues, as well as nutritional status of children undergoing repeated LVP, require further study in the pediatric population.
1. Tito M, Gines P, Arroyo V, et al. Total paracentesis associated with intravenous albumin in the management of patients with cirrhosis and ascites. Gastroenterology. 1990; 98: 146–51.
2. Angueira CE, Kadakia SC. Effects of large-volume paracentesis on pulmonary function in patients with tense cirrhotic ascites. Hepatology. 1994; 20: 825–8.
3. Byrd Jr, RP Roy TM, Simons M. Improvement in oxygenation after large volume paracentesis. South Med J. 1996; 89: 689–92.
4. Denzer BS. A new method of diagnosis of peritonitis in infancy and childhood. Am J Dis Child. 1920; 20: 113–5.
5. Kellett JW, Turner Jr, WR Levkoff AH. Paracentesis in the management of neonatal urinary ascites. Urology. 1973; 2: 672–5.
6. Baden HP, Morray JP. Drainage of tense ascites in children after cardiac surgery. J Cardiothorac Vasc Anesth. 1995; 9: 720–1.
7. Besson R, Gottrand F, Saulnier P, et al. Traumatic chylous ascites: conservative management. J Pediatr Surg. 1992; 27: 1543.
8. Unger SW, Chandler JG. Chylous ascites in infants and children. Surgery. 1983; 93: 455–61.
9. Gines P, Tito L, Arroyo V. Randomized comparative study of therapeutic paracentesis with and without intravenous albumin in cirrhosis. Gastroenterology. 1988; 94: 1493–502.
10. Planas R, Gines P, Arroyo V, et al. Dextran-70 versus albumin as plasma expanders in cirrhotic patients with tense ascites treated with total paracentesis. Results of a randomized study. Gastroenterology. 1990; 99: 1736–44.
11. Webster ST, Brown KL, Lucey MR, et al. Hemorrhagic complications of large volume abdominal paracentesis. Am J Gastroenterol. 1996; 91: 366–8.
12. Arnold C, Haag K, Blum HE, et al. Acute hemoperitoneum after large-volume paracentesis. Gastroenterology. 1997; 113: 978–82.
13. Wilcox CM, Woods BL, Mixon HT. Prospective evaluation of a peritoneal dialysis catheter system for large volume paracentesis. Am J Gastroenterol. 1992; 87: 1443–6.
14. Shaheen NJ, Grimm IS. Comparison of the Caldwell needle/canula with angiocath needle in large volume paracentesis. Am J Gastroenterol. 1996; 91: 1731–3.
© 2001 Lippincott Williams & Wilkins, Inc.