Retrospective Evaluation of Dosing Effects of Bumetanide Continuous Infusions in the Pediatric Cardiac Intensive Care Unit

Study Design.

This was a single center, descriptive, retrospective study performed at the University of Florida (UF) Health Shands Hospital in Gainesville, FL. Our PCICU unit is a quaternary care center with 23 beds for patients with complex congenital heart disease. Electronic health records were used to identify pediatric patients admitted to the PCICU who received a bumetanide CI.

Inclusion criteria were met if patients were younger than 6 years, admitted to the PCICU, and received bumetanide CI for at least 6 hours as part of their clinical management between January 1, 2018, and September 1, 2022. The age cutoff 6 years was used to allow for a more homogenous patient population and to evaluate the dosing effects in younger children, because previous studies that included pediatric patients up to 18 years of age have reported significant variations in dosing regimens.⁵ Patients with incomplete or missing medical records were excluded from the study. If a patient had multiple admissions during the study time frame, each encounter with bumetanide CI was assessed as an individual occurrence.

Data Collection.

A comprehensive review of medical records was conducted. Historical patient data were collected by manual chart review, and bumetanide CI was determined by clinical documentation on the medication administration record. Data collection included relevant demographic information (i.e., age, sex, weight); admitting cardiac problem or diagnosis; dosing regimens, including doses and duration of CI; previous use of furosemide; concomitant diuretics; fluid balance measurements (every 12 hours); weight measurements obtained 24 hours before and after bumetanide infusion and laboratory parameters (i.e., serum electrolytes; renal function biomarkers, namely blood urea nitrogen and serum creatinine every 12 and 24 hours); hemodynamic or mechanical circulatory support; and respiratory support and oxygen requirements, defined as any support other than room air to mechanical ventilatory and circulatory support requirements. Inadequate response to furosemide is defined as not reaching negative fluid balance.

Outcomes.

The primary outcome was identifying doses and total duration of bumetanide CI regimens. Secondary efficacy outcomes were determined by the ability to achieve negative fluid balance within 24 hours and time to reach negative fluid balance defined as the net cumulative balance of inputs and outputs with variables including but not limited to oral/IV fluids, medications, blood products, daily urine output, stool, and other bodily output (e.g., chest tube, gastrostomy, gastric or ostomy outputs) as recorded after initiation of bumetanide infusion. Secondary safety outcomes were based on the prevalence of electrolyte imbalances and renal impairment. Electrolyte imbalances were predefined as serum potassium concentration less than 3 mEq/L, serum chloride concentration less than 90 mEq/L, and serum bicarbonate concentration greater than 35 mEq/L. Renal impairment was predefined as increased serum creatinine (SCr) by 0.3 mg/dL within 48 hours or 1.5 times the baseline, or urine output <0.5 ml/kg/hr for at least 6 hours.^7,8 Laboratory values were obtained upon admission (or within 1 week of life for neonates) and within 12 hours of bumetanide peak infusion rates.

Subgroup analyses were performed to evaluate bumetanide regimens, based on patients’ overall cardiovascular (CV) history and risk. This evaluation included patients who required extracorporeal membrane oxygenation (ECMO) and those with an implanted ventricular assist device (VAD). Exploratory outcomes in the post hoc analysis included survival to discharge, mortality rates observed from peak bumetanide infusion rates, and dosing regimens used at the time of death.

Statistical Analysis.

Descriptive statistics (e.g., medians, IQRs, and frequencies) were used to summarize the study population’s patient demographic and clinical characteristics. The Wilcoxon signed rank test was used to test the statistical significance of continuous variables, and a chi-square test was used to analyze categorical data. All tests were 2-tailed, with an overall alpha level of 0.05 for statistical significance. To generate these findings, statistical analyses were performed by using appropriate statistical software, JMP Pro 17.

Baseline Characteristics.

Ninety-four patients were identified for study eligibility; however, 4 patients were excluded because they received continuous bumetanide infusions for less than 6 hours. Ninety patients were included in the final analysis; however, to account for patients having multiple hospital admissions, the total number of 106 hospital encounters was analyzed (Figure). Baseline patient demographics for all encounters, including age, sex, weight, and admitting cardiac problems, are summarized in Table 1. Most patients included were female and younger than 8 months (65 patients making up 72%). Eighty-six percent of the patients receiving bumetanide CI transitioned from furosemide with inadequate response to furosemide as defined by no negative fluid balance (Supplemental Table S1). Of note, 91% of patients received concomitant diuretics during bumetanide CI, with the highest use seen with chlorothiazide, followed by acetazolamide and spironolactone (Supplemental Figure). The most common admitting cardiac problem was hypoplastic left heart syndrome (HLHS), followed by cardiomyopathy and atrioventricular septal defect (AVSD).

View Figure

• Download as Powerpoint
• Download Figure

Table 1.Baseline Patient Demographics for All Encounters (N = 106)^*

View Fullscreen

Primary Outcome.

Table 2 highlights bumetanide dosing regimens observed in the PCICU during the study period. The median initial starting dose of bumetanide was 0.03 mg/kg/hr, ranging from 0.005 to 0.2 mg/kg/hr. The median dose of bumetanide was 0.046 mg/kg/hr, with a median duration of CI of 5.8 days. Median maximum doses of bumetanide were 0.075 mg/kg/hr, with doses as high as 0.3 mg/kg/hr being used in 2 patients.

Table 2.Bumetanide Dosing Regimen^*

View Fullscreen

Efficacy Outcomes.

The secondary efficacy endpoints for optimal diuresis are displayed in Table 3, with 83% of patients achieving negative fluid balance within 24 hours of bumetanide CI initiation and a median time to reach negative fluid balance of 13.4 hours. The urine output was collected for 24 hours for our study population with a median of 1.36 mL/kg/hr. To account for inconsistent or incomplete documentation on 18 encounters concerning weight, a weight comparative analysis was performed on 88 individuals before and 24 hours after starting a bumetanide CI. Before bumetanide initiation, the median weight was 5.0 kg compared with 4.9 kg after the start of bumetanide, which was statistically significant (p = 0.009).

Table 3.Efficacy Secondary Endpoints (N = 106)^*

View Fullscreen

Safety Outcomes.

Safety was analyzed by using the previously defined criteria for electrolyte imbalances and renal impairment. Seventy-three percent of patients experienced hypokalemia, 32% experienced hypochloremia, and 36% experienced hypercarbia (Table 4). Based on SCr elevations, 51% of patients had some renal impairment. Significant changes in serum electrolytes and renal function biomarkers were noted between all laboratory values measured at baseline and during median peak infusion rates but were not clinically significant (Table 5).

Table 4.Prevalence of Safety Secondary Endpoints (N = 106)^*

View Fullscreen

Table 5.Comparison Between Laboratory Values at Baseline and During Peak Infusion Rates^*

View Fullscreen

Exploratory Outcomes.

Epinephrine (45.2%), milrinone (34.4%), dopamine (8.6%), vasopressin (7%), norepinephrine (3.2%), and phenylephrine (1.6%) were the choices of vasopressor and inotropic support that required hemodynamic support for hemodynamic stability. Supplemental Table S2 includes additional clinical characteristics that were analyzed to assess patients’ CV history and risk. Eighteen percent of individuals were supported while on venoarterial (VA) ECMO, whereas 19% of patients had an implanted VAD using the Berlin Heart (Woodlands, TX, USA) EXCOR for additional mechanical circulatory support. When looking at heart transplant status, 34% of patients were awaiting heart transplants, whereas 12% had a history of heart transplants. Seventeen patients were supported by dialysis: 9 patients needing continuous renal replacement therapy (CRRT) and 8 patients requiring peritoneal dialysis (PD). Of the 17 patients, 13 were non-survivors, 8 received CRRT, and 5 were on PD. Of the 4 survivors, 1 was on CRRT, and 3 were on PD. Respiratory support and oxygen requirements varied substantially, with 75% of patients needing mechanical ventilation at some point during a bumetanide CI.

When assessing survival to discharge, 31 non-survivors were reported at the time of discharge. Of the non-survivors, 61% were receiving maximum bumetanide doses of 0.1 mg/kg/hr or greater (Supplemental Table S3). When stratifying this group further, based on maximum bumetanide doses (<0.1 vs ≥0.1 mg/kg/hr), there was a significant increase in deaths recorded with doses ≥0.1 mg/kg/hr (12 vs 19 patients; p = 0.0085). Doses observed at the time of death were 0.1 mg/kg/hr, ranging up to 0.3 mg/kg/hr (Supplemental Table S3).

A subgroup analysis of bumetanide regimens and survival to discharge, based on the CV history and risk, is presented in Supplemental Table S4. Patients with HLHS received higher doses of bumetanide and had higher mortality than those with cardiomyopathy and AVSD. Patients with a VAD in place received higher doses and had higher mortality rates than those without a VAD. This finding was also consistent among individuals on VA ECMO. Because several patients had multiple admissions, we analyzed these data from the number of encounters. For patients with multiple encounters, compared with those with a single encounter, we observed shorter durations of bumetanide CI (92 vs 161 hours; p = 0.3032) and increased survival to discharge (78% vs 68%; p = 0.3527). However, these results did not reach statistical significance.