Prospective observational study conducted in the pediatric emergency department (PED) of a tertiary care hospital in Spain between January 2022 and April 2023. The study was approved by the regional Research Ethics Committee.

Patients aged 3 months to 13 years with mild to moderate dehydration (assessed by means of the Gorelick23 scale or measured weight loss) secondary to AGE in whom ORT was contraindicated3 (losses greater than 10mL/kg/h, dehydration >10%, altered level of consciousness, severe deterioration of general health, paralytic ileus, acute abdomen or severe vomiting) or had failed. We sought informed consent from the legal guardians of the patients and patients aged 12 or more years.

We excluded patients in whom RIR was contraindicated1 (hemodynamic instability, serum sodium <130mmol/L or >150mmol/L, blood glucose >140mg/dL, systemic disease affecting hemodynamic and/or fluid and electrolyte homeostasis).

Patients with blood glucose levels of less than 50mg/dL who met all other inclusion criteria received a bolus of saline with 10% dextrose at a dose of 3mL/kg before administration of RIR and were included in the study if the capillary blood glucose level normalized within 10min, with the reading at this time recorded as the baseline glucose value.

We calculated that we needed a minimum sample size of 80 patients per group for a level of significance of 5%, a power of 80% and detection of a minimum mean difference of 1.5mmol/L in bicarbonate levels.

Patients who met the inclusion criteria based on a clinical evaluation and performance of laboratory tests (complete blood count, blood chemistry panel including renal function electrolyte panels, venous blood gas analysis, capillary ketone and glucose levels) received a course of RIR consisting of intravenous infusion over two hours of 40mL/kg (maximum of 1400mL) of one of the two crystalloid solutions compared in the study, selected based on the judgment of the physician in charge of the patient. Both solutions were supplemented with 2.5% dextrose through the addition of 25mL of 50% dextrose stock solution to 475mL of the crystalloid solution with the aim of correcting the increased ketone levels commonly found in these patients.1 The crystalloid solutions were prepared by the nursing staff of the PED per the customary protocol.

Patients remained in the observation unit of the PED during treatment. Once the course of RIR was completed, the patient underwent a second clinical evaluation and laboratory workup, in addition to assessment of oral tolerance with oral rehydration solution per the customary protocol. The physician in charge decided whether to discharge the patient home following a period of observation or admit the patient to the ward. Patients who received RIR at night remained in observation until the next morning to promote rest.

We collected data on demographic, anthropometric and clinical characteristics: age, sex, weight, symptoms, duration of symptoms, vital signs (heart rate, blood pressure, score in the Gorelick scale23) at diagnosis and after completion of RIR. The laboratory parameters recorded at baseline were urea, creatinine, C-reactive protein, electrolytes, venous blood gas, capillary ketone and glucose. The parameters recorded upon completion of RIR were venous blood gas and capillary ketone and glucose levels.

We documented the most common complications (venous infusion extravasation, signs of fluid overload) as well as any other unexpected adverse event to assess the safety of the RIR protocol.

All the data were collected prospectively in a form designed for the purpose, kept in the custody of the principal investigators, and eventually entered in a database developed with the SPSS software (IBM, Armonk, USA) following anonymization through the use of the numerical code assigned to each patient at the time of inclusion in the study.

The primary outcome was the change in bicarbonate values, expressed in mmol/L, during the patient’s participation in the study (bicarbonate after completion of RIR – baseline bicarbonate).

We analyzed the following secondary outcomes: change in venous chloride, sodium and potassium levels and in capillary glucose and ketone levels between baseline and RIR completion, length of stay in PED, frequency of hospital admission and readmission and incidence of adverse events.

The statistical analysis was performed with the software package SPSS and performed by intention to treat. We expressed normally distributed variables as mean and standard deviation, using the Student t test to compare groups. For data that did not follow a normal distribution, we used the median and interquartile range in the descriptive analysis and compared groups by means of the Mann-Whitney U test. Qualitative variables were expressed as absolute and relative frequencies and group comparisons made by means of the χ2 test.

For the assessment of changes in variables between baseline and completion of treatment, we used simple linear regression in the case of quantitative variables and univariate logistic regression models for qualitative variables.

The incidence of adverse events was compared using the χ2 test.

We considered results with a P value of less than 0.05 statistically significant.

A total of 186 patients were included in the study, of who 169 completed it; the patients had ages ranging from 6 to 167 months (median, 40 meses) and 49.1% were female; 86 (50.9%) received normal saline with 0.9% sodium chloride and 2.5% dextrose (D2.5NS) and the remaining 83 (49,1%) Plasma-Lyte with 2.5% dextrose (PLA with dextrose). Fig. 1 presents a flow diagram of the patients’ progress throughout the study.

Figure 1.

Flow diagram showing the progress of patients throughout the study.

(0.24MB).

Table 2 compares the demographic, anthropometric, clinical and laboratory characteristics of the patients at baseline. We did not find statistically significant differences between groups in any of these variables.

After completion of the course of RIR, patients in the PLA group showed greater recovery of serum bicarbonate levels with a lesser increase in serum chloride levels, compared to the D2.5NS group (Table 3). There were no significant differences between groups in the changes in sodium and potassium levels, and none of the patients developed hyperkalemia.

As regards blood glucose, of the 186 participants, 23 (12.4%) had hypoglycemia on arrival to the PED, requiring administration of a dextrose bolus prior to initiation of RIR. The mean ketone concentration in the sample at baseline was 4.2mmol/L, and levels had decreased after completion of IVR. We found an increase in blood glucose levels in the sample overall, with a greater increase in patients in the D2.5NS group (Table 3). However, the proportion of patients with a final glucose level greater than 140mg/dL was similar in both groups (36% of total sample; 30.1% of PLA with dextrose group, 41.8% of D2.5NS group; P=.149), as was the proportion of patients with final glucose levels greater than 200mg/dL (13.6% of total sample; 9.6% PLA with dextrose group; 17.4% of D2.5NS group; P=.144).

Compared to patient with normal glucose levels, patients who had developed hyperglycemia at the end of the study did not differ significantly in baseline glucose or ketone levels or in the proportion that received a dextrose bolus to correct hypoglycemia prior to initiation of RIR. However, patients with a final glucose level greater than 140mg/dL were significantly older compared to patients with final glucose levels in the normal range (Table 4).

The score in the Gorelick scale after RIR improved by nearly 2 points in the total sample.

Of the 161 patients who presented with vomiting on arrival, 3 received antiemetic treatment with ondansetron.

The mean length of stay was 19.2h. Eighteen percent of patients stayed in the PED for less than 6h. Eleven (6.5%) patients returned to the PED for the same reason, one of whom received a new course of RIR. As Table 5 shows, there were no significant differences in these parameters between the treatment groups.

The only observed adverse events were extravasation of the venous catheter in 4 of the 169 patients (2.4%), 2 in each treatment group, and development of edema in one patient who received D2.5NS course, with no statistically significant differences between groups.

We found improvement in blood gas analysis values in the entire sample, with an increase in serum bicarbonate levels, the laboratory parameter that is most strongly associated with dehydration in patients with AGE,24,25 corroborating the effectiveness of RIR therapy in these patients.

In the group of patients treated with a balanced crystalloid solution, reproducing the findings reported in previous studies,10 the increase in bicarbonate levels was significantly larger. In addition, serum chloride levels, a factor that contributes to the development or persistence of metabolic acidosis,13 were lower in this group, an outcome that can be attributed to the difference in the chloride concentration of the two crystalloid solutions.

Regarding the possible effects of fluid therapy on serum sodium and potassium levels, in agreement with previous studies,10 we found no differences in the changes of either ion levels based on the type of solution used for RIR. A recent Cochrane review11 suggests that the use of balanced solutions may reduce the incidence of hypokalemia after correction of the dehydration secondary to AGE. However, in our study, none of the patients developed hypo- or hyperkalemia during the follow-up. There were also no differences in the changes in potassium levels between the groups, evincing the safety of administering PLA in this respect, given its low potassium content (5mEq/L), the duration of RIR and the range of administered volumes (maximum of 40mL/kg), which corresponds to a maximum delivery of 0.2mEq per kg of body weight.

The entire sample exhibited a reduction in ketone levels of more than 2mmol/L after completion of RIR, consistent with the previous literature,26 with no differences between the groups, which was expected given that both solutions had the same dextrose content. This amount of dextrose supplementation seemed sufficient for the purposes of rehydration, contributing to the correction of high ketone levels, a factor involved in the persistence of vomiting.27 In fact, only 3 of the 161 patients who presented with vomiting on arrival required administration of ondansetron, which was reserved for patients in whom vomiting persisted in spite of bowel rest and IVR.

On the other hand, 36% of the sample had blood glucose levels greater than 140mg/dL upon completion of RIR, with capillary glucose values greater than 200mg/dL in more than 13% of patients, independently of baseline glucose and ketone levels, which brings into question whether the routine addition of 2.5% dextrose suggested in the guideline published by the Working Group on Hydration and Electrolyte Disorders of the Sociedad Española de Urgencias Pediátricas,1 even for patients with normal glucose levels, may not be excessive.

In our sample, patients with final blood glucose levels greater than 140mg/dL were older than patients with normal final blood glucose levels. This suggests that greater caution may be required in regard to dextrose supplementation as patient age increases.

Despite the marked differences found in the changes of biochemical parameters, the study did not identify significant differences in clinical outcomes in relation to the type of fluid administered to the patients.

We found improvement in the clinical assessment of hydration status by means of the Gorelick scale in the entire sample after administration of RIR, with no significant differences between the two groups. This was not consistent with the greater increase in bicarbonate levels observed in patients in the PLA group, although several publications question the usefulness of clinical scales for assessing the degree of dehydration in children with AEG.28,29

The mean length of stay in the total sample was 19h, slightly less than the 24h required for conventional IVR and shorter compared to the length reported in other studies,9,22 with no differences between the two treatment groups.

Although the main advantage of RIR approaches is a rapid clinical recovery allowing early discharge,6 in our case series only 18% of patients were discharged within 6h. However, it is difficult to interpret these findings given that the decision to discharge a patient home was not based on objective criteria, since the PED lacked a specific protocol for it, but rather left to the judgment of the physician in charge at the time. In addition, all patients that completed RIR at night were kept in observation until the morning after to allow adequate rest, which may have prolonged their stay in the PED beyond what was required based on their clinical condition.

Of the total sample, 6.5% of patients made return visits to the PED for the same complaint, with no differences between the two groups.

In agreement with the findings of previous studies,30 we did not identify any severe adverse event during the study beyond intravenous infusion extravasation in 4 of the 169 patients and development of self-limiting edema in one other patient. We did not identify any safety concerns associated with the differences in the composition of each of the solutions.