Peritoneal Dialysis vs Furosemide for Prevention of Fluid Overload in Infants After Cardiac Surgery: A Randomized Clinical Trial | Acid Base, Electrolytes, Fluids | JAMA Pediatrics | JAMA Network
[Skip to Navigation]
Sign In
Figure.  CONSORT Flow Diagram
CONSORT Flow Diagram

CONSORT flow diagram depicts study enrollment, completion, and withdrawals. CICU indicates cardiac intensive care unit; CPB, cardiopulmonary bypass; ECMO, extracorporeal circulation membrane oxygenation; OR, operating room; PD, peritoneal dialysis; and PDC, PD catheter.

Table 1.  Baseline Characteristics
Baseline Characteristics
Table 2.  Surgical Procedures by Group
Surgical Procedures by Group
Table 3.  Fluid Balance and Clinical Outcomes
Fluid Balance and Clinical Outcomes
1.
Bouchard  J, Soroko  SB, Chertow  GM,  et al; Program to Improve Care in Acute Renal Disease (PICARD) Study Group.  Fluid accumulation, survival and recovery of kidney function in critically ill patients with acute kidney injury.  Kidney Int. 2009;76(4):422-427.PubMedGoogle ScholarCrossref
2.
Sutherland  SM, Zappitelli  M, Alexander  SR,  et al.  Fluid overload and mortality in children receiving continuous renal replacement therapy: the prospective pediatric continuous renal replacement therapy registry.  Am J Kidney Dis. 2010;55(2):316-325.PubMedGoogle ScholarCrossref
3.
Stein  A, de Souza  LV, Belettini  CR,  et al.  Fluid overload and changes in serum creatinine after cardiac surgery: predictors of mortality and longer intensive care stay: a prospective cohort study.  Crit Care. 2012;16(3):R99.PubMedGoogle ScholarCrossref
4.
Hazle  MA, Gajarski  RJ, Yu  S, Donohue  J, Blatt  NB.  Fluid overload in infants following congenital heart surgery.  Pediatr Crit Care Med. 2013;14(1):44-49.PubMedGoogle ScholarCrossref
5.
Seguin  J, Albright  B, Vertullo  L,  et al.  Extent, risk factors, and outcome of fluid overload after pediatric heart surgery.  Crit Care Med. 2014;42(12):2591-2599.PubMedGoogle ScholarCrossref
6.
Hassinger  AB, Wald  EL, Goodman  DM.  Early postoperative fluid overload precedes acute kidney injury and is associated with higher morbidity in pediatric cardiac surgery patients.  Pediatr Crit Care Med. 2014;15(2):131-138.PubMedGoogle ScholarCrossref
7.
Bojan  M, Gioanni  S, Vouhé  PR, Journois  D, Pouard  P.  Early initiation of peritoneal dialysis in neonates and infants with acute kidney injury following cardiac surgery is associated with a significant decrease in mortality.  Kidney Int. 2012;82(4):474-481.PubMedGoogle ScholarCrossref
8.
Elahi  MM, Lim  MY, Joseph  RN, Dhannapuneni  RR, Spyt  TJ.  Early hemofiltration improves survival in post-cardiotomy patients with acute renal failure.  Eur J Cardiothorac Surg. 2004;26(5):1027-1031.PubMedGoogle ScholarCrossref
9.
Sorof  JM, Stromberg  D, Brewer  ED, Feltes  TF, Fraser  CD  Jr.  Early initiation of peritoneal dialysis after surgical repair of congenital heart disease.  Pediatr Nephrol. 1999;13(8):641-645.PubMedGoogle ScholarCrossref
10.
Shiao  CC, Wu  VC, Li  WY,  et al; National Taiwan University Surgical Intensive Care Unit-Associated Renal Failure Study Group.  Late initiation of renal replacement therapy is associated with worse outcomes in acute kidney injury after major abdominal surgery.  Crit Care. 2009;13(5):R171.PubMedGoogle ScholarCrossref
11.
Sanchez-de-Toledo  J, Perez-Ortiz  A, Gil  L,  et al.  Early initiation of renal replacement therapy in pediatric heart surgery is associated with lower mortality.  Pediatr Cardiol. 2016;37(4):623-628.PubMedGoogle Scholar
12.
Sasser  WC, Dabal  RJ, Askenazi  DJ,  et al.  Prophylactic peritoneal dialysis following cardiopulmonary bypass in children is associated with decreased inflammation and improved clinical outcomes.  Congenit Heart Dis. 2014;9(2):106-115.PubMedGoogle ScholarCrossref
13.
Kwiatkowski  DM, Menon  S, Krawczeski  CD,  et al.  Improved outcomes with peritoneal dialysis catheter placement after cardiopulmonary bypass in infants.  J Thorac Cardiovasc Surg. 2015;149(1):230-236.PubMedGoogle ScholarCrossref
14.
Alkan  T, Akçevin  A, Türkoglu  H,  et al.  Postoperative prophylactic peritoneal dialysis in neonates and infants after complex congenital cardiac surgery.  ASAIO J. 2006;52(6):693-697.PubMedGoogle ScholarCrossref
15.
Jenkins  KJ.  Risk adjustment for congenital heart surgery: the RACHS-1 method.  Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2004;7:180-184.PubMedGoogle ScholarCrossref
16.
Hershberger  SL. Hodges-Lehmann estimators. In: Lovric  M, ed.  International Encyclopedia of Statistical Science. Berlin, Germany: Springer-Verlag; 2011:635-636.
17.
Sampaio  TZ, O’Hearn  K, Reddy  D, Menon  K.  The influence of fluid overload on the length of mechanical ventilation in pediatric congenital heart surgery.  Pediatr Cardiol. 2015;36(8):1692-1699.PubMedGoogle ScholarCrossref
18.
Arikan  AA, Zappitelli  M, Goldstein  SL, Naipaul  A, Jefferson  LS, Loftis  LL.  Fluid overload is associated with impaired oxygenation and morbidity in critically ill children.  Pediatr Crit Care Med. 2012;13(3):253-258.PubMedGoogle ScholarCrossref
19.
Averbuch  N, Birk  E, Frenkel  G,  et al.  Percutaneous intraperitoneal catheters in neonates following open heart surgery.  J Intensive Care Med. 2014;29(3):160-164.PubMedGoogle ScholarCrossref
20.
Marenzi  G, Lauri  G, Grazi  M, Assanelli  E, Campodonico  J, Agostoni  P.  Circulatory response to fluid overload removal by extracorporeal ultrafiltration in refractory congestive heart failure.  J Am Coll Cardiol. 2001;38(4):963-968.PubMedGoogle ScholarCrossref
21.
Dries  DJ, McGonigal  MD, Malian  MS, Bor  BJ, Sullivan  C.  Protocol-driven ventilator weaning reduces use of mechanical ventilation, rate of early reintubation, and ventilator-associated pneumonia.  J Trauma. 2004;56(5):943-951.PubMedGoogle ScholarCrossref
22.
Alghamdi  AA, Singh  SK, Hamilton  BC,  et al.  Early extubation after pediatric cardiac surgery: systematic review, meta-analysis, and evidence-based recommendations.  J Card Surg. 2010;25(5):586-595.PubMedGoogle ScholarCrossref
23.
Cheng  DC, Karski  J, Peniston  C,  et al.  Early tracheal extubation after coronary artery bypass graft surgery reduces costs and improves resource use: a prospective, randomized, controlled trial.  Anesthesiology. 1996;85(6):1300-1310.PubMedGoogle ScholarCrossref
24.
Vincent  JL, Bihari  DJ, Suter  PM,  et al; EPIC International Advisory Committee.  The prevalence of nosocomial infection in intensive care units in Europe: results of the European Prevalence of Infection in Intensive Care (EPIC) Study.  JAMA. 1995;274(8):639-644.PubMedGoogle ScholarCrossref
25.
Raymond  TT, Cunnyngham  CB, Thompson  MT, Thomas  JA, Dalton  HJ, Nadkarni  VM; American Heart Association National Registry of CPR Investigators.  Outcomes among neonates, infants, and children after extracorporeal cardiopulmonary resuscitation for refractory inhospital pediatric cardiac arrest: a report from the National Registry of Cardiopulmonary Resuscitation.  Pediatr Crit Care Med. 2010;11(3):362-371.PubMedGoogle Scholar
26.
Macdonald  JE, Struthers  AD.  What is the optimal serum potassium level in cardiovascular patients?  J Am Coll Cardiol. 2004;43(2):155-161.PubMedGoogle ScholarCrossref
27.
Hodgkin  JE, Soeprono  FF, Chan  DM.  Incidence of metabolic alkalemia in hospitalized patients.  Crit Care Med. 1980;8(12):725-728.PubMedGoogle ScholarCrossref
28.
Bagshaw  SM, Delaney  A, Haase  M, Ghali  WA, Bellomo  R.  Loop diuretics in the management of acute renal failure: a systematic review and meta-analysis.  Crit Care Resusc. 2007;9(1):60-68.PubMedGoogle Scholar
29.
Ho  KM, Power  BM.  Benefits and risks of furosemide in acute kidney injury.  Anaesthesia. 2010;65(3):283-293.PubMedGoogle ScholarCrossref
30.
Ryerson  LM, Mackie  AS, Atallah  J,  et al.  Prophylactic peritoneal dialysis catheter does not decrease time to achieve a negative fluid balance after the Norwood procedure: a randomized controlled trial.  J Thorac Cardiovasc Surg. 2015;149(1):222-228.PubMedGoogle ScholarCrossref
31.
Wauters  J, Claus  P, Brosens  N, McLaughlin  M, Malbrain  M, Wilmer  A.  Pathophysiology of renal hemodynamics and renal cortical microcirculation in a porcine model of elevated intra-abdominal pressure.  J Trauma. 2009;66(3):713-719.PubMedGoogle ScholarCrossref
32.
Dittrich  S, Aktuerk  D, Seitz  S,  et al.  Effects of ultrafiltration and peritoneal dialysis on proinflammatory cytokines during cardiopulmonary bypass surgery in newborns and infants.  Eur J Cardiothorac Surg. 2004;25(6):935-940.PubMedGoogle ScholarCrossref
33.
Valeri  A, Radhakrishnan  J, Vernocchi  L, Carmichael  LD, Stern  L.  The epidemiology of peritonitis in acute peritoneal dialysis: a comparison between open- and closed-drainage systems.  Am J Kidney Dis. 1993;21(3):300-309.PubMedGoogle ScholarCrossref
34.
Swan  P, Darwish  A, Elbarbary  M, Al Halees  Z.  The safety of peritoneal drainage and dialysis after cardiopulmonary bypass in children.  J Thorac Cardiovasc Surg. 1997;114(4):688-689.PubMedGoogle ScholarCrossref
35.
Murala  JS, Singappuli  K, Provenzano  SC  Jr, Nunn  G.  Techniques of inserting peritoneal dialysis catheters in neonates and infants undergoing open heart surgery.  J Thorac Cardiovasc Surg. 2010;139(2):503-505.PubMedGoogle ScholarCrossref
36.
Krawczeski  CD, Goldstein  SL, Woo  JG,  et al.  Temporal relationship and predictive value of urinary acute kidney injury biomarkers after pediatric cardiopulmonary bypass.  J Am Coll Cardiol. 2011;58(22):2301-2309.PubMedGoogle ScholarCrossref
Original Investigation
April 2017

Peritoneal Dialysis vs Furosemide for Prevention of Fluid Overload in Infants After Cardiac Surgery: A Randomized Clinical Trial

Author Affiliations
  • 1Heart Institute, Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio
  • 2Division of Cardiology, Department of Pediatrics, Stanford University, Palo Alto, California
  • 3Center for Acute Care Nephrology, Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio
JAMA Pediatr. 2017;171(4):357-364. doi:10.1001/jamapediatrics.2016.4538
Key Points

Question  In infants who undergo cardiac surgery, is peritoneal dialysis associated with improved fluid balance compared with furosemide treatment?

Findings  In this randomized clinical trial that included 73 infants, patients receiving peritoneal dialysis had a lower incidence of fluid overload and no significant adverse outcomes compared with patients receiving furosemide.

Meaning  Use of peritoneal dialysis after cardiac surgery is safe and associated with a lower incidence of fluid overload than is furosemide treatment among infants at high risk for postoperative acute kidney injury.

Abstract

Importance  Fluid overload after congenital heart surgery is frequent and a major cause of morbidity and mortality among infants. Many programs have adopted the use of peritoneal dialysis (PD) for fluid management; however, its benefits compared with those of traditional diuretic administration are unknown.

Objective  To determine whether infants randomized to PD vs furosemide for the treatment of oliguria have a higher incidence of negative fluid balance on postoperative day 1, as well as avoidance of 10% fluid overload; shorter duration of mechanical ventilation, intensive care unit stay, and inotrope use; and fewer electrolyte abnormalities.

Design, Setting, and Participants  This single-center, unblinded, randomized clinical trial compared methods of fluid removal after cardiac surgery from October 1, 2011, through March 13, 2015, in a large tertiary pediatric hospital in Ohio. The parents or guardians of all eligible infants (aged <6 months) undergoing cardiac surgery with catheter placement for PD were approached for inclusion. No patients were withdrawn for adverse effects. Recruitment was powered for the primary outcome, and analysis was based on intention to treat. Patients randomized to PD were hypothesized to have superior outcomes.

Interventions  Infants received intravenous furosemide (1 mg/kg every 6 hours) or a standardized PD regimen.

Main Outcomes and Measures  The primary end point was incidence of negative fluid balance on postoperative day 1. Secondary end points included incidence of fluid overload, duration of mechanical ventilation and intensive care unit stay, electrolyte abnormalities and repletion doses, duration of inotropic administration, and mortality.

Results  Seventy-three patients (47 boys [64%] and 26 girls [35%]; median age, 8 [interquartile range (IQR), 6-14] days) received treatment and completed the trial. No difference was found between the PD and furosemide groups in the incidence of negative fluid balance on the first postoperative day. The furosemide group was 3 times more likely to have 10% fluid overload (odds ratio [OR], 3.0; 95% CI, 1.3-6.9), was more likely to have prolonged ventilator use (OR, 3.1; 95% CI, 1.2-8.2), and had a longer duration of inotrope use (median, 5.5 [IQR, 4-8] vs 4.0 [IQR, 3-6] days) and higher electrolyte abnormality scores (median, 6 [IQR, 4-7] vs 3 [IQR, 2-5]) compared with the PD group. No statistically significant differences in mortality (3 patients [9.4%] in the furosemide group vs 1 patient [3.1%] in the PD group) or length of cardiac intensive care unit (median, 7 [IQR, 6-12] vs 9 [IQR, 5-15] days) or hospital (15 [IQR, 10-28] vs 14 [IQR, 9-22] days) stay were observed. No serious complications were observed. Dialysis was discontinued early in 9 of 41 patients in the PD group for pleural-peritoneal communication.

Conclusions and Relevance  Use of PD is safe and allows for superior fluid management with improved clinical outcomes compared with diuretic administration. Use of PD should be strongly considered among infants at high risk for postoperative acute kidney injury and fluid overload.

Trial Registration  clinicaltrials.gov Identifer: NCT01709227

Introduction

Infants are at high risk for fluid overload after cardiac surgery secondary to impaired hemodynamics, acute kidney injury (AKI), and capillary leak. Fluid overload is commonly self-limited and often marginalized as an expected stage in recovery. However, studies have shown that overload is associated with worse outcomes in critically ill patients.1-3 Among infants after cardiac surgery, fluid overload independently predicts mortality and morbidity.4-6

The paradigm of early renal replacement therapy has gained attention as studies have demonstrated associations with lower mortality and improved outcomes after cardiac surgery.7-11 Among infants, peritoneal dialysis (PD) is the most common modality of dialysis, proven to be a safe method of fluid removal.7,9,12-14 A previously published retrospective study13 demonstrated that PD is associated with improved outcomes, including duration of mechanical ventilation. However, to our knowledge, PD has not been compared prospectively with diuretic administration, the traditional postoperative therapy in the context of oliguria.

In this randomized clinical trial, we aimed to determine whether the modality of fluid removal (PD vs a standardized furosemide regimen) is associated with fluid balance and clinical outcomes. We hypothesized that, compared with infants receiving furosemide, infants randomized to PD would be more likely to have a negative fluid balance on postoperative day (POD) 1, would be less likely to develop 10% fluid overload, would have less time to negative fluid balance, and would have superior clinical outcomes, including duration of mechanical ventilation, length of stay, vasoactive infusion use, electrolyte level abnormalities, oxygenation indices, and mortality.

Methods

This single-center, randomized clinical trial among infants with oliguria after cardiac surgery with cardiopulmonary bypass (CPB) compared patients undergoing PD vs furosemide administration. Parents or guardians of all infants (aged <6 months) undergoing cardiac surgery with CPB at Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, with planned PD catheter (PDC) placement as part of routine practice were approached for enrollment. Patients with preexisting kidney disease (estimated glomerular filtration rate, <60 mL/min/1.73m2) were excluded. Patients were removed from the study before randomization if they did not undergo CPB or if a PDC was not placed as planned or after randomization if the patient died or required a second operation or extracorporeal membrane oxygenation within the first postoperative 24 hours, although their data were collected for the safety analysis. Patient screening was performed by review of the cardiac intensive care unit (CICU) census and operating room schedule. The study protocol is available in Supplement 1. The study protocol was approved by the institutional review board of the Cincinnati Children’s Hospital with independent Data Safety Monitoring Board supervision. Written informed consent was obtained preoperatively from the parents or the guardians of patients.

PD Management

Per institutional protocol, a PDC was placed in patients deemed to be at high risk for AKI (eTable 1 in Supplement 2). Nephrology services were consulted preoperatively. The PD system was maintained as a closed circuit, replaced every 72 hours with minimal circuit manipulation by dialysis nurses.

After CPB, a 2-cm transverse midline incision was placed 2 cm below the sternotomy by the cardiac surgeon. The 37-cm catheter (Pediatric Tenckhoff; Quinton Instrument Company) was placed by perforating the peritoneum above the diaphragm via the sternotomy. A right-angle clamp was placed into the peritoneum and passed through the abdominal wall into the skin incision. The catheter was pulled into the peritoneum, positioned in the right lower quadrant, and sutured to the abdominal wall. The peritoneum was closed. The PDC was tested for leaks, connected to a closed system, and clamped until randomization or study completion.

Study Procedures

The surgical procedure, CPB, and perfusion were performed with the surgeon blinded to study inclusion. Thus, a decision to not place a PDC could occur regardless of enrollment.

Postoperative oliguria was defined as any 4 total hours of urine output of less than 1 mL/kg per hour during the first postoperative 24 hours. Infants who developed oliguria were randomized to PD or furosemide (Figure). Patients without oliguria received standard care without study intervention.

Patients randomized to PD had an initial regimen of PD solution, 1.5%, 10 mL/kg (Dianeal; Baxter Healthcare) with 2 to 3 mEq/L of potassium chloride and 200 U/L of unfractionated heparin. Standard initial peritoneal dialysis regimen was to fill the peritoneal cavity with the PD solution for 5 minutes, allow the solution to dwell for 45 minutes, then drain it passively for 10 minutes. This cycle was repeated for the duration of PD using a closed system setup (Gesco; Utah Medical Products). This regimen is our standard initial infant treatment. The percentage of glucose, dwell volumes, or cycle duration was adjusted with nephrology consultation.

Patients randomized to furosemide treatment were prescribed 1 mg/kg intravenously every 6 hours for 2 doses, then per physician direction. Chlorothiazide sodium was added on POD 2 at physician discretion. A urine output of less than 1 mL/kg per hour during the 16 hours after the first furosemide dose was considered to be a poor response, and these patients were allowed to initiate PD or undergo catheter drainage. If patients developed late fluid overload unresponsive to diuretic therapy, PD use was outside the study protocol.

All patients received inotrope and vasoactive medications as dictated by clinical status. Fluid administration was two-thirds maintenance (Holliday-Segar method) on POD 1, three-fourths maintenance on POD 2, and 100% maintenance thereafter as standard therapy. Patients received electrolyte supplementation and fluid boluses as needed.

Randomization was block stratified 1:1 by the risk adjustment for congenital heart surgery (RACHS-1) severity score using 2 strata (score 2-5 and score 6; range, 0-6, with higher scores indicating higher risk for inpatient mortality).15 The randomization scheme was created using a permuted block method with a random number generator with blocks of 2, 4, and 6 patients. The cardiology division statistician confidentially prepared envelopes and maintained the assignment log. After randomization, the statistician was blinded to the study. Medical records were interrogated by clinical research coordinators from the Heart Institute Research Core, Cincinnati Children’s Hospital Medical Center, and transferred to a REDCap database using double-data entry with appropriate data cleaning.

Study Variables

Baseline intraoperative variables included surgical procedure and RACHS-1 score, CPB and aortic cross-clamp times, and use of deep hypothermic circulatory arrest, regional-cerebral perfusion, or modified ultrafiltration. The primary outcome data consisted of the percentage of patients with a negative fluid balance on POD 1, calculated as the net difference between all inputs and outputs. Secondary outcomes included the rate of 10% fluid overload (fluid balance divided by preoperative weight, expressed as a percentage), number of 8-hour shifts until negative fluid balance was achieved, duration of mechanical ventilation to the first extubation and vasoactive infusions (in days), electrolyte repletion doses, morning electrolyte abnormality score (range, 0-4, with higher scores indicating more abnormalities)13 from PODs 1 to 5, modified oxygenation index13 and level of brain-type natriuretic peptide at postoperative 24 and 48 hours, CICU and hospital lengths of stay, days to chest closure if applicable, and in-hospital mortality (eTable 2 in Supplement 2). Prolonged mechanical ventilation was defined post hoc as longer than 3 days and prolonged CICU stay was defined as longer than 7 days.

Incidence of AKI was not calculated because PD modulates postoperative creatinine and urine output. Furthermore, PD use defines AKI by established guidelines.

Study Withdrawals

Randomized patients who did not complete their assigned treatment were replaced with the next eligible patient with an identical RACHS-1 score after completion of standard enrollment. Analysis was performed using all patients in an intention-to-treat format and with those completing protocol treatment. For clarity, we report intention to treat with per protocol treatment in eTables 3 and 4 in Supplement 2.

Statistical Analysis

Demographic and clinical characteristics were summarized using measures of central tendency, variability, and frequency. Appropriate statistical tests were applied to describe treatment groups and compare the characteristics. Outcomes were analyzed using the χ2 test, Fisher exact test, 2-sample t test, or Wilcoxon rank sum test as appropriate. Difference in medians of continuous variables was determined using the Hodge-Lehmann estimator16 with 95% CIs calculated by the exact distribution of the Mann-Whitney test statistic. All tests were 2-sided, with P < .05 considered to be statistically significant. Data were analyzed using SPSS software (version 23.0; IBM SPSS Statistics).

Power calculations were performed on the primary outcome, initially using conservative estimates, but recalculated using a prospectively planned adaptive study design with the outcomes of the first 20 enrolled patients. Among these patients (10 in each group), 9 (90%) in the PD group and 6 (60%) in the furosemide group had negative fluid balance. Using a 2-group χ2 test with a 2-sided significance level of .05, 32 patients per group were required for 80% power to detect difference (90% vs 60%; odds ratio [OR], 6.0; 95% CI, 0.5-67.7) (nQuery Advisor, version 7.0; Statistical Solutions).

Results

Enrollment lasted from October 1, 2011, through March 13, 2015. Of 119 eligible patients, 102 consented for inclusion (Figure). Twenty-nine patients were removed per protocol or did not meet randomization criteria. Of the original 32 patients randomized to PD, 6 did not undergo PD owing to a pleural-peritoneal communication. This occurrence resulted in 9 replacement patients (3 replacements also had communications), leading to 41 patients randomized to PD and 32 randomized to furosemide (73 randomized patients; 47 boys [64%] and 26 girls [35%]; median age, 8 [interquartile range (IQR), 6-14] days).

We found no differences in baseline or demographic data, including median weight, age, and baseline creatinine level (Table 1). Cohorts had similar RACHS-1 scores, duration of CPB and aortic cross-clamping, use of regional-cerebral perfusion, deep hypothermic circulatory arrest, and modified ultrafiltration. Represented surgical procedures are listed in Table 2. Mean (SD) volume of fluid administered included 98 (26) vs 104 (23) mL/kg per day on POD 1, 108 (30) vs 116 (31) mL/kg per day on POD 2, 111 (30) vs 116 (26) mL/kg per day on POD 3, 115 (30) vs 121 (32) mL/kg per day on POD 4, and 125 (29) vs 123 (31) mL/kg per day on POD 5. No difference in nutritional volume administered was found.

Patients randomized to furosemide treatment received a median (IQR) daily dose of 3 (3-4), 4 (3-4), 3 (3-4), 3 (3-4), and 3 (3-4) mg/kg on PODs 1 to 5, respectively. Nine patients (28%) received a peak daily dose of more than 4 mg/kg per day. Similarly, 15 patients (47%) were given chlorothiazide on POD 2 and 16 (50%) on POD 3. Eight patients in the furosemide group had inadequate diuretic response after 24 hours and had their PDC opened to drain (4 patients) or used for dialysis (4 patients), which was initiated in 1 patient before the second furosemide dose (Figure).

We found no statistically significant difference in the percentage of patients who attained a negative fluid balance on POD 1 (29 of 41 [71%] vs 21 of 32 [66%]; P = .80) (Table 3). However, patients randomized to furosemide were 3 times more likely to develop 10% fluid overload (OR, 3.0; 95% CI, 1.3-6.9). In addition, although not statistically significant, the PD cohort achieved a net negative fluid balance 1 shift sooner and had superior median fluid balances on POD 1 (55 [IQR, −25 to 194] vs 118 [IQR, 51 to 197] mL), POD 2 (−251 [IQR, −379 to −125] vs −201 [IQR, −343 to −100] mL), and POD 3 (−89 [IQR, −210 to −14] vs −66 [IQR, −193 to 54] mL).

Patients randomized to furosemide treatment were more likely to require prolonged mechanical ventilation (OR, 3.1; 95% CI, 1.2-8.2), although without a statistically significant difference in the median duration of mechanical ventilation (3 [IQR, 2-4] vs 4 [IQR, 20-6] days). Patients in the PD group were less likely to have prolonged CICU stays (OR, 1.6; 95% CI, 1.0-2.7) but had a nonsignificant difference in actual duration of CICU stay (median, 7 [IQR, 6-12] vs 9 [IQR, 5-15] days). Dichotomous outcomes were defined post hoc. Patients treated with PD had fewer electrolyte repletion doses (median, 1 [IQR, 0-2] vs 2 [IQR, 1-5]) and lower electrolyte abnormality scores (median, 3 [IQR, 2-5] vs 6 [IQR, 4-7]). The PD cohort had a 1.5-day shorter duration of vasoactive infusion use (median, 4.0 [IQR, 3-6] vs 5.5 [IQR, 4-8] days). Eighteen patients had delayed sternal closure with a median of 2.5 (IQR, 2-3) days per cohort. In-hospital mortality included 4 patients (5.5%) among all randomized patients, with no statistically significant difference (3 patients [9.4%] in the furosemide group vs 1 [2.4%] in the PD group).

Among all 92 patients with PDC placement, regardless of study intervention, 2 PD-related adverse outcomes occurred. Neither included peritonitis or bowel injury. One nonrandomized patient had self-resolving bloody drainage from their PDC (in addition to typical postsurgical bleeding) and received a blood transfusion. This catheter was later used for effective PD outside the study protocol. One patient had a patent processus vaginalis and developed a hydrocele that resolved after PD was completed, requiring no intervention.

Pleural-peritoneal communications were present in 9 of 41 patients in the PD group. This condition was diagnosed by drainage of dialysis fluid via chest tubes. This procedural complication is the correlate of pericatheter leak in patients with a direct transcutaneous PDC placement and was more frequent in patients weighing less than 3 kg (4 of 8 [50%] vs 5 of 33 [15%]; OR, 5.6; 95% CI, 1.0-30.1). These patients did not use PD as assigned; however, the complications were not considered to be adverse events because no harm occurred.

Discussion

Prevention of fluid overload in critically ill children is one of few modifiable risk factors in postoperative management. As evidence of the perils of fluid overload mounts, the conversation must shift toward treatment and prevention. This study reveals that early PD is a superior method of fluid management with a very low risk for adverse events and is associated with less-prolonged mechanical ventilation, fewer inotropic requirements, and fewer electrolyte abnormalities.

We did not find a statistically significant difference in the percentage of patients attaining negative fluid balance on POD 1. This result may be attributable to fluid removal limitations related to hemodynamic instability and ongoing capillary leak as well as timing of PD initiation, often within POD 1, which limits the net effects in this time frame. However, patients randomized to PD were less likely to develop fluid overload and had improved fluid management, including time to negative fluid balance and daily fluid balance.

Our data build on previous studies. Postoperative fluid overload has been associated with prolonged mechanical ventilation, more inotropic requirements, longer hospital stay, and increased mortality.2,4,6,17 Positive fluid balance is associated with impaired gas exchange and reduced lung compliance.18 Drainage of free abdominal fluid increases oxygenation.19 Cardiac output and stroke volume are diminished in fluid overload and improve with ultrafiltration.20

Patients randomized to PD were less likely to require prolonged duration of mechanical ventilation. Avoidance of mechanical ventilation may decrease the incidence of ventilator-associated pneumonia, use of sedative and paralytic agents, and overall use of resources.21-23 Ventilator-associated pneumonia is the most common nosocomial infection in the intensive care unit, associated with a doubling of mortality.24 Furthermore, infants with congenital heart disease often have improved hemodynamics with spontaneous ventilation. Of note, prolonged ventilation was a post hoc definition, and this association must be interpreted cautiously and confirmed in future studies.

Peritoneal dialysis facilitated more fluid removal while also achieving improved electrolyte levels. Electrolyte balance is important for optimal cardiac function and prevention of serious arrhythmias. Imbalance is a risk factor for death among critically ill infants.25 Hypokalemia may increase the risk for ventricular arrhythmias and sudden cardiac death and cause diastolic dysfunction.26 Diuretic-induced metabolic alkalosis decreases respiratory drive and delays ventilator weaning.27

In a previous matched-cohort study of infants with intraoperative PDC placement,13 patients undergoing PD had improved fluid balance, earlier extubation, improved inotrope scores, and fewer electrolyte level imbalances. Another cohort study12 showed association of prophylactic PD with negative fluid balance and decreased inotrope requirements compared with passive drainage. Similarly, a propensity score–guided study7 demonstrated that earlier PD was associated with decreased mortality among those requiring dialysis. Furosemide is used aggressively for oliguria in many CICU settings, despite studies refuting this practice among patients with AKI.28,29

Peritoneal dialysis faces skepticism fueled by underpowered studies. A randomized study of PD after Norwood palliations reported that PD was not associated with negative fluid balance, but more adverse effects.30 The authors reported that 4 of 10 patients randomized to PD had cardiac arrest; however, 3 had cardiac arrest before PD initiation. The PD group had worse preoperative cardiac function and lactate levels and almost half never underwent dialysis. Some critique PD for increased costs; however, PD expenditures are offset by decreased resource use resulting in unchanged costs.13

We did not find a mortality difference. Surgical advancements have made death rare, and thus interventions that may affect mortality require very large samples to reveal treatment effect. Similarly, we could not comment on time to delayed sternal closure with only 18 applicable patients. We are unable to assess renal protection, because dialysis clears creatinine and long-term follow-up is not yet available.

The mechanistic effects of PD are multifactorial. Peritoneal dialysis allows for direct drainage of extravascular fluid and a decrease in edema. This effect may improve venous and lymphatic drainage and increase renal perfusion pressure.31 Peritoneal dialysis may also modulate the cytokine milieu implicit in post-CPB capillary leak syndrome. Many investigators12,32 postulate that PD may remove maladaptive cytokines, specifically interleukins 6 and 8, thus limiting protein and large molecule extravasation and thereby preventing capillary leak. Cytokine clearance may also prevent nephron damage, improving renal clearance.32

We reported adverse outcomes in all 92 randomized and nonrandomized patients undergoing PDC placement. The 2 events were minor and did not affect outcomes. Of importance, no patient had peritonitis, bowel injury, or discontinuation of PD for hemodynamic instability, cited in older studies.9,13 Our safety profile is similar to that of most recent studies and largely attributed to the intraoperative catheter placement, use of a closed circuit replaced every 72 hours, low dwell volumes, and circuit care by trained dialysis nurses.9,33,34 Multiple catheter placement techniques have been described with varying complication rates.35 Our institution uses cuffed catheters placed by indirect transperitoneal methods with the cuff remaining outside the skin. Half of patients undergoing PD and weighing less than 3 kg were unable to receive dialysis owing to pleural-peritoneal communication, which likely involved incomplete maintenance of a watertight seal after a defect was created in the extremely thin peritoneum prone to leakage. Even if dialysis does not work initially, adhesion formation may allow functional catheter use several days later.

Our findings only slightly varied when including patients who had nonfunctioning catheters in the analysis. We believe that our results are most valid and generalizable using intention-to-treat analysis. In practice, the PDC may malfunction, a problem that needs to be considered in a risk-benefit analysis.

Limitations and Strengths

This study is, to our knowledge, the first randomized clinical trial to evaluate PD in infants and is notable for a relatively large size and regimented treatment arms. However, our study has limitations. Although the single-center design may limit generalizability, it allows control of confounding variables and was used with equipoise such that nearly all patients received the assigned treatment, avoiding withdrawal after randomization for personal preference. The starting furosemide dose (1 mg/kg every 6 hours) may be less than some would choose in particular scenarios but was chosen to avoid early postoperative hemodynamic instability. The protocol allowed dose increases after the first 2 doses. Patients randomized to the furosemide arm with inadequate diuresis were also allowed to receive dialysis at clinical discretion. Although this procedure may introduce bias, we believed this was ethically necessary, because all patients underwent PDC placement per our standard practice. Similarly, recruitment was limited to patients with normal baseline renal function because we believe that keeping a PD catheter clamped on a patient with baseline renal dysfunction and a high risk for AKI is ethically improper. Removal of these biases would only strengthen our findings. Last, our study design was potentially biased by not blinding physicians. However, all patients had PDCs, which somewhat limited this effect. Delaying randomized therapy until oliguria also potentially limited treatment effects.

This study was not powered to determine a difference in clinical outcomes. Although not statistically significant, differences in ventilation and CICU days between groups carry clinical importance and potentially statistical significance in larger cohorts. In subcategory analysis, differences were not statistically significant, presumably owing to sample size.

Additional studies on the utility of PD can be strengthened by a multicentered design and use of novel biomarkers or genetic polymorphisms predictive of AKI36 in a biomarker-directed study. These adaptations would allow better confounder control and earlier initiation of treatment. Biomarker-guided recruitment would maximize treatment effect and study power because only high-risk patients would receive the intervention.

Conclusions

We reveal that PD is safe and allows superior fluid management with improved clinical outcomes compared with diuretic administration among infants with oliguria after cardiac surgery. Peritoneal dialysis use should be strongly considered among those at high risk for AKI and subsequent fluid overload. Our findings highlight the need for a larger, multicentered study to understand the benefits of PD.

Back to top
Article Information

Corresponding Author: David M. Kwiatkowski, MD, MS, Division of Cardiology, Department of Pediatrics, Stanford University, 750 Welch Rd, Ste 321, Palo Alto, CA 94304 (david.kwiatkowski@stanford.edu).

Accepted for Publication: November 20, 2016.

Published Online: February 20, 2017. doi:10.1001/jamapediatrics.2016.4538

Author Contributions: Dr Kwiatkowski had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kwiatkowski, Goldstein, Nelson, Morales, Krawczeski.

Acquisition, analysis, or interpretation of data: Kwiatkowski, Goldstein, Cooper, Nelson, Krawczeski.

Drafting of the manuscript: Kwiatkowski, Goldstein, Morales,

Krawczeski.

Statistical analysis: Kwiatkowski, Goldstein

Obtained funding: Kwiatkowski.

Administrative, technical, or material support: Kwiatkowski, Cooper, Krawczeski.

Study supervision: Kwiatkowski, Goldstein, Nelson, Morales, Krawczeski.

Conflict of Interest Disclosures: Dr Goldstein reports consulting, speaker’s bureau, and grant support from Baxter International/Gambro Renal Products and consulting for Akebia Therapeutics, AM-Pharma BV, Bellco Health Corporation, La Jolla Pharmaceutical Company, and Otsuka Pharmaceutical Company. No other disclosures were reported.

Funding/Support: This study was supported by grant 13POST16240002 from the American Heart Association–Great Rivers Affiliate and internal funding from Cincinnati Children’s Hospital Medical Center.

Role of the Funder/Sponsor: The funders had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

References
1.
Bouchard  J, Soroko  SB, Chertow  GM,  et al; Program to Improve Care in Acute Renal Disease (PICARD) Study Group.  Fluid accumulation, survival and recovery of kidney function in critically ill patients with acute kidney injury.  Kidney Int. 2009;76(4):422-427.PubMedGoogle ScholarCrossref
2.
Sutherland  SM, Zappitelli  M, Alexander  SR,  et al.  Fluid overload and mortality in children receiving continuous renal replacement therapy: the prospective pediatric continuous renal replacement therapy registry.  Am J Kidney Dis. 2010;55(2):316-325.PubMedGoogle ScholarCrossref
3.
Stein  A, de Souza  LV, Belettini  CR,  et al.  Fluid overload and changes in serum creatinine after cardiac surgery: predictors of mortality and longer intensive care stay: a prospective cohort study.  Crit Care. 2012;16(3):R99.PubMedGoogle ScholarCrossref
4.
Hazle  MA, Gajarski  RJ, Yu  S, Donohue  J, Blatt  NB.  Fluid overload in infants following congenital heart surgery.  Pediatr Crit Care Med. 2013;14(1):44-49.PubMedGoogle ScholarCrossref
5.
Seguin  J, Albright  B, Vertullo  L,  et al.  Extent, risk factors, and outcome of fluid overload after pediatric heart surgery.  Crit Care Med. 2014;42(12):2591-2599.PubMedGoogle ScholarCrossref
6.
Hassinger  AB, Wald  EL, Goodman  DM.  Early postoperative fluid overload precedes acute kidney injury and is associated with higher morbidity in pediatric cardiac surgery patients.  Pediatr Crit Care Med. 2014;15(2):131-138.PubMedGoogle ScholarCrossref
7.
Bojan  M, Gioanni  S, Vouhé  PR, Journois  D, Pouard  P.  Early initiation of peritoneal dialysis in neonates and infants with acute kidney injury following cardiac surgery is associated with a significant decrease in mortality.  Kidney Int. 2012;82(4):474-481.PubMedGoogle ScholarCrossref
8.
Elahi  MM, Lim  MY, Joseph  RN, Dhannapuneni  RR, Spyt  TJ.  Early hemofiltration improves survival in post-cardiotomy patients with acute renal failure.  Eur J Cardiothorac Surg. 2004;26(5):1027-1031.PubMedGoogle ScholarCrossref
9.
Sorof  JM, Stromberg  D, Brewer  ED, Feltes  TF, Fraser  CD  Jr.  Early initiation of peritoneal dialysis after surgical repair of congenital heart disease.  Pediatr Nephrol. 1999;13(8):641-645.PubMedGoogle ScholarCrossref
10.
Shiao  CC, Wu  VC, Li  WY,  et al; National Taiwan University Surgical Intensive Care Unit-Associated Renal Failure Study Group.  Late initiation of renal replacement therapy is associated with worse outcomes in acute kidney injury after major abdominal surgery.  Crit Care. 2009;13(5):R171.PubMedGoogle ScholarCrossref
11.
Sanchez-de-Toledo  J, Perez-Ortiz  A, Gil  L,  et al.  Early initiation of renal replacement therapy in pediatric heart surgery is associated with lower mortality.  Pediatr Cardiol. 2016;37(4):623-628.PubMedGoogle Scholar
12.
Sasser  WC, Dabal  RJ, Askenazi  DJ,  et al.  Prophylactic peritoneal dialysis following cardiopulmonary bypass in children is associated with decreased inflammation and improved clinical outcomes.  Congenit Heart Dis. 2014;9(2):106-115.PubMedGoogle ScholarCrossref
13.
Kwiatkowski  DM, Menon  S, Krawczeski  CD,  et al.  Improved outcomes with peritoneal dialysis catheter placement after cardiopulmonary bypass in infants.  J Thorac Cardiovasc Surg. 2015;149(1):230-236.PubMedGoogle ScholarCrossref
14.
Alkan  T, Akçevin  A, Türkoglu  H,  et al.  Postoperative prophylactic peritoneal dialysis in neonates and infants after complex congenital cardiac surgery.  ASAIO J. 2006;52(6):693-697.PubMedGoogle ScholarCrossref
15.
Jenkins  KJ.  Risk adjustment for congenital heart surgery: the RACHS-1 method.  Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2004;7:180-184.PubMedGoogle ScholarCrossref
16.
Hershberger  SL. Hodges-Lehmann estimators. In: Lovric  M, ed.  International Encyclopedia of Statistical Science. Berlin, Germany: Springer-Verlag; 2011:635-636.
17.
Sampaio  TZ, O’Hearn  K, Reddy  D, Menon  K.  The influence of fluid overload on the length of mechanical ventilation in pediatric congenital heart surgery.  Pediatr Cardiol. 2015;36(8):1692-1699.PubMedGoogle ScholarCrossref
18.
Arikan  AA, Zappitelli  M, Goldstein  SL, Naipaul  A, Jefferson  LS, Loftis  LL.  Fluid overload is associated with impaired oxygenation and morbidity in critically ill children.  Pediatr Crit Care Med. 2012;13(3):253-258.PubMedGoogle ScholarCrossref
19.
Averbuch  N, Birk  E, Frenkel  G,  et al.  Percutaneous intraperitoneal catheters in neonates following open heart surgery.  J Intensive Care Med. 2014;29(3):160-164.PubMedGoogle ScholarCrossref
20.
Marenzi  G, Lauri  G, Grazi  M, Assanelli  E, Campodonico  J, Agostoni  P.  Circulatory response to fluid overload removal by extracorporeal ultrafiltration in refractory congestive heart failure.  J Am Coll Cardiol. 2001;38(4):963-968.PubMedGoogle ScholarCrossref
21.
Dries  DJ, McGonigal  MD, Malian  MS, Bor  BJ, Sullivan  C.  Protocol-driven ventilator weaning reduces use of mechanical ventilation, rate of early reintubation, and ventilator-associated pneumonia.  J Trauma. 2004;56(5):943-951.PubMedGoogle ScholarCrossref
22.
Alghamdi  AA, Singh  SK, Hamilton  BC,  et al.  Early extubation after pediatric cardiac surgery: systematic review, meta-analysis, and evidence-based recommendations.  J Card Surg. 2010;25(5):586-595.PubMedGoogle ScholarCrossref
23.
Cheng  DC, Karski  J, Peniston  C,  et al.  Early tracheal extubation after coronary artery bypass graft surgery reduces costs and improves resource use: a prospective, randomized, controlled trial.  Anesthesiology. 1996;85(6):1300-1310.PubMedGoogle ScholarCrossref
24.
Vincent  JL, Bihari  DJ, Suter  PM,  et al; EPIC International Advisory Committee.  The prevalence of nosocomial infection in intensive care units in Europe: results of the European Prevalence of Infection in Intensive Care (EPIC) Study.  JAMA. 1995;274(8):639-644.PubMedGoogle ScholarCrossref
25.
Raymond  TT, Cunnyngham  CB, Thompson  MT, Thomas  JA, Dalton  HJ, Nadkarni  VM; American Heart Association National Registry of CPR Investigators.  Outcomes among neonates, infants, and children after extracorporeal cardiopulmonary resuscitation for refractory inhospital pediatric cardiac arrest: a report from the National Registry of Cardiopulmonary Resuscitation.  Pediatr Crit Care Med. 2010;11(3):362-371.PubMedGoogle Scholar
26.
Macdonald  JE, Struthers  AD.  What is the optimal serum potassium level in cardiovascular patients?  J Am Coll Cardiol. 2004;43(2):155-161.PubMedGoogle ScholarCrossref
27.
Hodgkin  JE, Soeprono  FF, Chan  DM.  Incidence of metabolic alkalemia in hospitalized patients.  Crit Care Med. 1980;8(12):725-728.PubMedGoogle ScholarCrossref
28.
Bagshaw  SM, Delaney  A, Haase  M, Ghali  WA, Bellomo  R.  Loop diuretics in the management of acute renal failure: a systematic review and meta-analysis.  Crit Care Resusc. 2007;9(1):60-68.PubMedGoogle Scholar
29.
Ho  KM, Power  BM.  Benefits and risks of furosemide in acute kidney injury.  Anaesthesia. 2010;65(3):283-293.PubMedGoogle ScholarCrossref
30.
Ryerson  LM, Mackie  AS, Atallah  J,  et al.  Prophylactic peritoneal dialysis catheter does not decrease time to achieve a negative fluid balance after the Norwood procedure: a randomized controlled trial.  J Thorac Cardiovasc Surg. 2015;149(1):222-228.PubMedGoogle ScholarCrossref
31.
Wauters  J, Claus  P, Brosens  N, McLaughlin  M, Malbrain  M, Wilmer  A.  Pathophysiology of renal hemodynamics and renal cortical microcirculation in a porcine model of elevated intra-abdominal pressure.  J Trauma. 2009;66(3):713-719.PubMedGoogle ScholarCrossref
32.
Dittrich  S, Aktuerk  D, Seitz  S,  et al.  Effects of ultrafiltration and peritoneal dialysis on proinflammatory cytokines during cardiopulmonary bypass surgery in newborns and infants.  Eur J Cardiothorac Surg. 2004;25(6):935-940.PubMedGoogle ScholarCrossref
33.
Valeri  A, Radhakrishnan  J, Vernocchi  L, Carmichael  LD, Stern  L.  The epidemiology of peritonitis in acute peritoneal dialysis: a comparison between open- and closed-drainage systems.  Am J Kidney Dis. 1993;21(3):300-309.PubMedGoogle ScholarCrossref
34.
Swan  P, Darwish  A, Elbarbary  M, Al Halees  Z.  The safety of peritoneal drainage and dialysis after cardiopulmonary bypass in children.  J Thorac Cardiovasc Surg. 1997;114(4):688-689.PubMedGoogle ScholarCrossref
35.
Murala  JS, Singappuli  K, Provenzano  SC  Jr, Nunn  G.  Techniques of inserting peritoneal dialysis catheters in neonates and infants undergoing open heart surgery.  J Thorac Cardiovasc Surg. 2010;139(2):503-505.PubMedGoogle ScholarCrossref
36.
Krawczeski  CD, Goldstein  SL, Woo  JG,  et al.  Temporal relationship and predictive value of urinary acute kidney injury biomarkers after pediatric cardiopulmonary bypass.  J Am Coll Cardiol. 2011;58(22):2301-2309.PubMedGoogle ScholarCrossref
×