Evaluation of the Two-Bag System for Fluid Management in Pediatric Patients with Diabetic Ketoacidosis

INTRODUCTION

In the United States, diabetic ketoacidosis (DKA) accounts for nearly 15% of all pediatric admissions each year.1 It also accounts for approximately 20% of all mortality in children due to diabetes mellitus (DM).2 Some factors that need to be considered when managing children with DKA include electrolyte imbalance, fluid loss, and hyperglycemia.3 One of the immediate and vital interventions initiated is fluid replacement therapy. The purpose of this therapy is to replace fluid loss and correct electrolyte imbalances.

Frequent intravenous (IV) fluid modification is necessary to adapt to ongoing changes in fluid balance, serum glucose, and electrolytes.1 A onebag and a two-bag system have been used for fluid management of pediatric DKA patients. The one-bag system, however, has been noted to have limitations, such as slow response time and increase in hospital cost.4 The two-bag system uses two bags of fluids containing the same amount of electrolytes but different concentrations of dextrose. By adjusting the rates of each bag, clinicians can immediately provide a customized glucose infusion rate while keeping the fluid delivery constant (Figure).1 The two-bag system has been utilized since the 1990s.4 Over the years, small clinical studies have shown few advantages to using the two-bag system, such as faster fluid exchange rate, less IV fluid bags used and reduced cost of IV fluid therapy.14 None of the studies, however, have shown any clinical benefit in terms of rate of glucose or bicarbonate correction.

View Figure

• Download as Powerpoint
• Download Figure

The purpose of our study was to assess whether the two-bag system conveys any clinical benefit compared to the one-bag system. We also assessed what fluids were used at Moses Cone Hospital for fluid management in pediatric DKA patients.

MATERIALS AND METHODS

RESULTS

Thirty-one pediatric patients were included in this study. The one-bag system was used in 9 patients (29%) and the two-bag system was used in 22 patients (71%). Baseline characteristics were similar between the two groups, with the exception of baseline serum creatinine (SCr) and initial insulin drip rate (Table 1). The patients in the one-bag system group had a higher baseline SCr (p < 0.001), while the patients in the two-bag system group had a higher initial insulin drip rate (p < 0.001). The mean age of the two groups was 13.5 years (p = 0.384). The mean weight was also similar between groups (two-bag: 49.5 ± 16.4 kg; one-bag: 55.4 ± 17.7 kg; p = 0.475). The majority of the patients were female (61%).

Table 1. Baseline Characteristics of Study Population

View Fullscreen

On admission, the mean hemoglobin A_1c of the subjects in the two-bag system group was 12.4 ± 1.9% and that in the one-bag system group was 10.7 ± 2.4% (p = 0.753). Thirty-nine percent (n = 12) of the population had a history of DKA and 71% (n = 22) had a history of DM. Of the 22 patients who had a history of DM, all were receiving insulin therapy prior to admission. Only 3 patients (10%) were taking oral diabetes medications, with metformin being the most common.

The mean initial glucose level in the two-bag group was 489.5 ± 134.4 mg/dL; and in the onebag group, 435.2 ± 205.8 mg/dL (p = 0.171). The patients in the two-bag group presented with an anion gap of 18.0 ± 4.8; the anion gap in the onebag group was 16.6 ± 6.7 (p = 0.121). The mean initial pH in the two-bag group was 7.166 ± 0.096 and that in the one-bag group was 7.186 ± 0.128 (p = 0.344). Other initial laboratories were similar between groups.

The severity of DKA was similar between groups (p = 0.55). DKA is usually classified into three major categories: mild (venous pH < 7.3 or bicarbonate < 15 mEq/L), moderate (venous pH < 7.2 or bicarbonate < 10 mEq/L), and severe (venous pH < 7.1 or bicarbonate < 5 mEq/L).5 In this study, the majority of the patients (n=13) fell into the mild category; 11 (35%) were in the moderate category.

The mean duration of fluids infused was similar between groups (two-bag, 31.5 ± 18.6 hour; one-bag, 31.7 ± 17.4 hour; p = 0.845). The majority of the patients received an insulin drip. The mean duration of insulin drip in the two-bag group was 16.6 ± 10.3 hour and in the one-bag group was 15.0 ± 5.6 hour (p = 0.106).

Types of Potassium Used

The IV fluids used in both the two-bag and one-bag group contained potassium phosphate or potassium chloride or both (Table 2). Potassium acetate was used only 18% of the time in the two-bag group and 22% in the one-bag group.

Table 2. Types of Potassium Used

View Fullscreen

Rate of CBG and Bicarbonate Correction

The mean ± SD rate of CBG correction was 31.04 ± 20.61 mg/dL/hr in the two-bag group and 21.04 ± 16.26 mg/dL/hr in the one-bag group. This difference was not significant (p = 0.297). The two-bag group had a faster rate of bicarbonate correction as compared to the one-bag group (0.949 ± 0.553 mEq/L/hr and 0.606 ± 0.297 mEq/L/hr, respectively; p = 0.047) (Table 3).

Table 3. Clinical Results of Using Two-Bag vs. One-Bag System

View Fullscreen

DISCUSSIONS

Despite the small sample size, our results suggest that the two-bag system can provide a faster rate of bicarbonate and ketone correction in pediatric patients with diabetic ketoacidosis. We also found a trend towards a faster rate of CBG and pH correction with the two-bag system. The latter endpoints were not significant probably due to the small sample size and the retrospective nature of this study. The rate of CBG and bicarbonate correction were used as the primary endpoints in this study because these two parameters were also used in a previous study to assess the efficacy of the two-bag system.1 The time to pH and ketone correction were used as secondary endpoints because they potentially can serve as indicators of metabolic acidosis resolution and lipolysis cessation, which are big components in the pathophysiology of diabetic ketoacidosis.5

The mean rate of CBG correction with the two-bag system in this study was similar to that reported in the prospective study published by Poirier and colleagues.1 In their study, the mean (95% CI) rate of CBG correction with the twobag system was 33.1 mg/dL/hr (95% CI, 22–44 mg/dL/hr) compared to 31.04 mg/dL/hr in our study. The mean rate of CBG correction with the one-bag system was quite different (current study: 21.04 vs Poirier: 30.2 mg/dL/hr). This difference may be explained by the small sample size of the one-bag group in our study (n = 9) as compared to the study by Poirier and colleagues (n = 16). Also, their study was randomized and prospective.1 According to the literature,13 the goal rate of CBG decline with insulin drip should not be faster than 100 mg/dL/hr due to the concern of hypoglycemia. This concern is another reason that dextrose is administered to the patients once their CBG reaches 250–300 mg/dL.5 A rapid glucose correction can theoretically cause a sudden shift in serum osmolarity, which may lead to extrapontine myelinolysis.6 However, the rate of glucose correction has not been found to increase the risk of cerebral edema or brainstem herniation.78

Since both continued ketone production and reduced bicarbonate concentration can lead to elevated anion gap, more rapid bicarbonate and ketone correction achieved by the two-bag system can potentially resolve the metabolic acidosis faster.3 These endpoints, in turn, may decrease days of pediatric intensive care and total hospital stay; however, these potential benefits have not been proven in clinical studies.

Correcting acidosis faster can ultimately prevent the use of bicarbonate therapy. Bicarbonate therapy in DKA patients may increase hepatic ketone production and lead to paradoxical acidosis of the cerebrospinal fluid and serum hypertonicity (due to hypokalemia and hypernatremia).35 Correcting acidosis with bicarbonate therapy has also been associated with an increased risk of cerebral edema.7

Even though the patients in the one-bag system group had a higher baseline SCr, this observation is probably not clinically significant in this particular study. However, the difference in the initial insulin drip rate between the two groups may be clinically significant because insulin resistance can be offset by an insulin concentration of around 100–200 μU/mL, which can induce cessation of lipolysis and ketogenesis. This concentration can usually be achieved by an insulin drip rate of around 0.1 units/kg/hr.5 Thus, the higher initial rate of the insulin drip in the two-bag group may have influenced some of the study endpoints.

No particular pattern was observed in the choice of IV fluids and electrolytes used in both the two-bag and the one-bag group. This inconsistency of fluid prescribing may indicate that more education for the house staff would be beneficial. Guidelines to help standardize the prescribing of IV fluid therapy for pediatric patients with DKA may also be valuable.

The two-bag system, as indicated by previously published studies, provides easier management of DKA patients by nursing staff. It provides a faster fluid exchange rate, fewer IV fluid bags used, and reduced cost of IV fluid therapy.14 Using one standard method of managing pediatric DKA patients within an institution could help reduce the potential for errors.

A limitation to this study is that the size of the two groups was not equal. This discrepancy is probably due to the fact that at our institution, most of the pediatric intensive care attending physicians prefer using the two-bag system for DKA fluid management.

In conclusion, using the two-bag system to manage pediatric patients with DKA in our institution appeared to provide a faster rate of bicarbonate and ketone correction. A larger, randomized prospective study to further look into this issue would be beneficial.