Background. The serum electrolyte panel (SEP) is a frequently ordered laboratory test, but it has unproven usefulness in the treatment of dehydrated pediatric patients. Our study purpose was to evaluate the usefulness of routinely ordering a SEP in the treatment of dehydrated pediatric patients receiving intravenous fluids (IVFs).
Methods. Children 2 months to 9 years of age who were receiving IVFs because of dehydration were prospectively studied in a pediatric emergency department (PED). Historical data, physical examination findings, degree of dehydration, and SEP results were recorded. After patient evaluation, attending physicians documented whether they would have ordered a SEP. Outcome measurements included changes in clinical management on the basis of SEP results, as well as correlations of dispositions and unscheduled return visits (URVs) with SEP results.
Results. A total of 182 patients were enrolled in the study. One hundred eleven patients had mild dehydration, 55 moderate dehydration, and 16 severe dehydration. Eighty-eight patients (48%; 95% confidence interval: 41–56%) had ≥1 abnormal SEP value. Clinically relevant findings included bicarbonate levels of <16 mmol/L for 28% of patients, hypoglycemia for 9.9%, hypokalemia for 6.0%, and hypernatremia for 3.0%. The attending physicians predicted that a SEP would be clinically important for 34% of all patients. There was a 58% sensitivity in detecting which children would have clinically significant SEP results. Overall, SEP results changed clinical management in 10.4% of cases. One hundred sixty-five (91%) of the patients were discharged from the PED (including 48 who were initially observed), of whom 7 (3.8%) had URVs to the PED within 72 hours and were given additional IVFs. Seventeen patients were admitted (median: 2.6 days), 2 of whom had URVs after hospital discharge for additional IVFs.
Conclusions. On the basis of initial presentation, attending physicians were poor at predicting which children would have clinically significant SEP results. Low bicarbonate values were correlated with observation unit use but not with hospitalization or URVs. The observation unit provided effective care for a subset of dehydrated patients, avoiding the need for hospitalization. Obtaining a SEP can provide useful information for the treatment of some children receiving IVFs because of dehydration.
Acute evaluation and treatment of children presenting with dehydration represent one of the most common situations in the pediatric emergency department (PED); yearly visits to the emergency department (ED) by children 0 to 18 years of age with dehydration averaged ∼384 500 for 1999–2000.1 Acute gastroenteritis, the most common cause of dehydration, ranks second to acute respiratory infections in overall pediatric disease frequency.2 Up to 10% of all US hospital admissions of children <5 years of age are because of diarrhea and dehydration.3
The serum electrolyte panel (SEP) is frequently ordered but has unproven usefulness in the care of dehydrated pediatric patients. Only 1 study examined the predictive value of the SEP, showing that, among children 6 months to 13 years of age with acute gastroenteritis who received intravenous fluids (IVFs) because of dehydration, those with serum bicarbonate levels of ≤13 mmol/L were more likely to be hospitalized.4 We found no literature pertaining to the usefulness of obtaining a SEP for all dehydrated pediatric patients receiving IVFs.
The purpose of this study was to evaluate the usefulness of routinely obtaining a SEP for patients 2 months to 9 years of age who received IVF rehydration for treatment of dehydration. Specifically, we set out to describe the frequency of SEP abnormalities, the frequency with which SEP abnormalities altered clinical care, the frequency with which physicians could predict SEP abnormalities, and the association between the SEP results and patient disposition.
During a 10-month period (January to October), we enrolled a convenience sample of patients, 2 months to 9 years of age, who presented to an urban PED with gastroenteritis and dehydration, all of whom received IVFs for rehydration. The need for IVF rehydration was determined by the attending physician, and treated children included children with persistent emesis or altered mental status and those who experienced failure of a trial of oral rehydration in the PED. Oral rehydration in our institution consists of small amounts (often administered with a syringe) of an oral rehydration solution (ie, Pedialyte, Ross Laboratories, Columbus, OH), offered every several minutes, or popsicles for older children. Failure to tolerate oral rehydration therapy consisted of ongoing vomiting, inability to take the oral rehydration fluids, or severe diarrhea, with water and electrolyte loss. Exclusion criteria included inborn errors of metabolism, cardiac disease, renal disease, or taking a medication with the potential to alter serum electrolyte levels.
Eligible patients were enrolled by the department's staff of 8 attending physicians (all board certified in pediatric emergency medicine). All attending physicians recorded their responses on a standardized data collection form. The attending physicians recorded the degree of dehydration before IVF rehydration with a standardized instrument with previously accepted clinical profiles of dehydration (Table 1). Every physical examination finding on the form was recorded for every patient. When children had findings for >1 of the dehydration categories (mild, moderate, or severe), the attending physicians assigned 1 overall value of estimated percent dehydration that was considered best correlated with the patient's perceived degree of dehydration. This final assigned percent dehydration value was used to place the patient into 1 of the 3 dehydration categories. In addition, physicians recorded age, length of illness, ounces of oral intake, episodes of emesis and diarrhea, and urine output (Table 2). After the initial evaluation and before laboratory data were obtained, attending physicians recorded whether they would have ordered a SEP and, if so, the indication for ordering this test. Blood (∼1.0 mL) was drawn at the time of intravenous (IV) catheter insertion for all patients and was sent to the laboratory for SEP analysis. After reviewing the SEP results, the attending physician recorded how the SEP results changed the medical treatment of the patient. Patient disposition was left to the discretion of the attending physicians but generally was based on rehydration indicated by physical examinations, the ability to tolerate oral fluids, and no ongoing significant fluid losses.
Telephone follow-up interviews were performed to assess whether patients had unscheduled return visits (URVs) to their physicians or to any other facility to obtain additional care or IVFs within the subsequent 72 hours. Medical records for all patients were later reviewed for URVs within 72 hours to our children's hospital. These charts were reviewed for assessment of the type of care rendered on the URV.
Primary outcome measures included the incidence of SEP abnormalities and how these abnormalities changed medical management. Secondary outcome measures included the physician's ability to predict SEP abnormalities and patient disposition.
Frequency of SEP Abnormalities
We first described the frequency of SEP abnormalities in our sample for each electrolyte in the panel, as well as an overall value for the complete panel. SEP results were considered abnormal if they were outside the age-adjusted ranges defined by our clinical laboratory. The following definitions of SEP abnormalities were used: sodium levels of <130 mmol/L or >150 mmol/L, bicarbonate levels of <16 mmol/L, potassium levels of <3.5 mmol/L, and glucose levels of <60 mg/dL. On the basis of the cutoff values used in previous studies,4 we chose to perform bivariate analysis by using a bicarbonate level of ≤13 mmol/L as an indication of a significantly abnormal SEP. The following prediction variables were then compared: age, percent dehydration, length of illness, oral intake, number of episodes of emesis, number of episodes of diarrhea, and number of urinations. Subanalyses included comparing the bicarbonate values for children <1 year of age versus >1 year of age, as well as correcting for differences in percent dehydration and the number of episodes of diarrhea.
How SEP Abnormalities Changed Medical Management
We measured the frequency of the attending physicians' responses that the SEP results changed their medical management of the case. SEP abnormalities were classified as clinically significant electrolyte abnormalities (CSEAs) if they changed the medical management. We report results for each abnormal electrolyte component of the SEP, and we detail how the management was reported to have changed or what action resulted after the SEP findings were reviewed.
Physicians' Decisions to Order a SEP and Sensitivity in Detecting a CSEA
Attending physicians were asked to document whether they would have ordered a SEP after completion of the history recording and physical examination, before the SEP results were available. A comparison was then made between this decision and the presence of a CSEA.
Disposition outcomes were categorized as follows: 1) discharged from the PED, 2) placed in the PED 6-bed observation unit for continued IVF administration, beginning 2 hours into the PED rehydration course, 3) admitted to the hospital because of dehydration, or 4) URV to the ED within 72 hours (either given additional IVFs and then discharged from the PED or given additional IVFs and then admitted). We first analyzed these disposition outcomes by comparing them with the percent dehydration at presentation. We grouped the physicians' estimates of percent dehydration into 3 categories, ie, mild (1–5%), moderate (6–10%), and severe (≥10%), and described the frequency of disposition within each category. We then compared serum bicarbonate levels of ≤13 mmol/L with disposition outcomes. We also performed logistic regression of the outcomes extended care (either placed in observation or admitted) and URV with the independent variable abnormal SEP.
Descriptive statistics, including proportions, means, and medians, are reported. Confidence intervals (CIs) (95%), including exact binomial CIs for proportions and CIs for differences in means, are reported when appropriate. Other statistical methods used included t tests (Fisher's exact test when appropriate) and Spearman correlations. Regression was used to examine the effects of age on bicarbonate levels, with adjustment for episodes of diarrhea. Odds ratios relating the unit change in a continuous or ordinal variable to a dichotomous outcome (eg, admission or observation) were calculated with logistic regression. Statistical analyses were performed with S-Plus 2000 Professional Release 3 (Mathsoft, Cambridge, MA).
The institutional review board approved this study. Informed consent was obtained from parents or legal guardians, and assent was obtained from children >7 years of age.
Seventy percent of all eligible patients were enrolled, resulting in 182 study participants (median age: 1.4 years; range: 2.7 months to 8.5 years). Thirty-two percent of the patients were <1 year of age, and 51% were male. Sixty-one percent (111 patients) were classified as having mild dehydration, 30% (55 patients) as having moderate dehydration, and 9% (16 patients) as having severe dehydration. Patient flow according to disposition outcome is shown in Fig 1. IVF administration was 42 mL/kg for patients treated in the ED and then discharged, compared with 103 mL/kg for patients placed in the observation unit (P < .0001). Eighty-one percent of the patients (147 of 182 patients) were contacted for telephone follow-up interviews. In addition, medical records for all patients were reviewed for repeat hospital visits.
Frequency of SEP Abnormalities
The overall incidence of SEP abnormalities was 48% (95% CI: 41–56%) for ≥1 value. Ninety-six percent of the patients (174 patients) presented with isonatremic dehydration (sodium levels of 130–150 mmol/L). One percent of the patients (2 patients) had serum sodium levels of 130 mmol/L, and no patients had hyponatremia (sodium levels of <130 mmol/L). Three percent of the patients (6 patients) had hypernatremia; 1 patient had a serum sodium level of >160 mmol/L, and the rest had values between 150 and 155 mmol/L. The patient with a serum sodium level of 168 mmol/L was a child with Dubowitz syndrome and severe dehydration. There were 11 children (6%) with hypokalemia, with values ranging from 1.5 to 3.4 mmol/L. The 1 significantly low value (1.5 mmol/L) occurred for a child with a history of being fed a homemade solution of water mixed with baking soda. The remaining children had mild hypokalemia, ranging from 2.8 to 3.4 mmol/L. Hypoglycemia occurred for 9.9% of the patients (18 of 182 patients), with a mean age of 18.4 months. These values ranged from 37 to 59 mg/dL. Elevated blood urea nitrogen (BUN) levels occurred for 28% of the patients (50 patients), with a range of 18 to 107 mg/dL and a mean of 24.4 mg/dL. Low serum bicarbonate levels (<16 mmol/L) represented the most common individual SEP abnormality, occurring for 29% of the study population. The 1 child with an elevated serum bicarbonate levels, mentioned above, had been fed sodium bicarbonate.
Patient characteristics are presented on the basis of serum bicarbonate levels of either ≤13 mmol/L or >13 mmol/L (Table 2). This serum bicarbonate level is a level that many clinicians consider to be clinically significant, as reported in at least 1 ED report.4 In our study, patients with serum bicarbonate levels of ≤13 mmol/L were more likely to be younger children, to have higher clinical estimated degrees of dehydration, to have had more ounces of oral intake, and to have had more episodes of diarrhea. In a comparison of children <1 year of age with those ≥1 year of age, children <1 year of age had a mean serum bicarbonate level that was lower by 2.0 mmol/L (95% CI: 0.7–3.3 mmol/L), despite similar mean degrees of dehydration (<1 year: 5.8%; ≥1 year: 6.1%). Additional analyses were performed for these 2 age groups, to determine whether the amount of diarrhea accounted for this difference. Mean episodes of diarrhea were 7.8 episodes (median: 6 episodes) for children <1 year of age, compared with 4.3 episodes (median: 2 episodes) for children ≥1 year of age (P < .0001). After adjustment for this difference, however, the serum bicarbonate level was still lower by 2.1 mmol/L (95% CI: 1.1-3.0 mmol/L) for children <1 year of age.
How SEP Abnormalities Changed Medical Management
The SEP abnormality was clinically significant, ie, the physician indicated it changed management, for 10.4% of the patients. How each abnormal SEP result changed medical management is shown in Table 3. Abnormal serum bicarbonate and glucose levels were the most common electrolyte values associated with changes in medical management. Although elevated BUN levels were common abnormalities, there was no reported change in management based on these elevated values.
Physicians' Decisions to Order a SEP and Sensitivity in Detecting a CSEA
After the initial evaluations, attending physicians indicated that they would have ordered a SEP in 34% of the cases. All attending physicians indicated that the decision to order a SEP was made before the results of the SEP were reviewed. Among the children for whom the attending physician would have ordered a SEP, there was a 58% sensitivity in detecting a CSEA (Table 4). The hypoglycemic patients were not identified with the history or physical examination.
Patient dispositions according to the initial degree of dehydration are illustrated in Fig 2. The majority of the patients discharged from the PED were assessed as being mildly dehydrated. Compared with the mildly dehydrated patients, the group with moderate dehydration had a higher percentage of patients placed in the observation unit. Most patients with an URV were assessed as mildly dehydrated. None of the severely dehydrated patients had an URV.
Disposition analyses according to individual SEP components focused on the serum sodium, bicarbonate, and glucose levels. Sodium levels of >150 mmol/L were not associated with observation room placement, hospital admission, or URVs. There were 53 patients with bicarbonate levels of <16 mmol/L; of those patients, 28 were discharged from the PED and 20 were placed in an observation unit, given additional IVFs (range: 2–15 hours after presentation; median: 8.8 hours), and then discharged. Five patients were admitted to the hospital. The mean serum bicarbonate level for children placed in the observation unit was 2.0 mmol/L (95% CI: 0.9–3.1 mmol/L) less than the mean serum bicarbonate level for children discharged. We then presented our disposition outcome data with a serum bicarbonate level of ≤13 mmol/L (Table 5). Children with serum bicarbonate levels of ≤13 mmol/L were more likely to undergo observation than were children with serum bicarbonate levels of >13 mmol/L. Otherwise, there was not a difference between the 2 groups. Interestingly, all 9 patients with URVs had serum bicarbonate levels of >13 mmol/L. Although not statistically significant, this finding showed a trend that was opposite that expected. There might be a variety of reasons why these patients with URVs returned for care. However, this group did not undergo additional analyses.
During examination of the overall correlation between extended stay (either admission or observation) and abnormal SEP results, 2 factors were noted. Each 10-fold increase in the BUN level increased the odds of admission/observation 7.1-fold (95% CI: 1.5–33.5-fold). Each decrease in serum bicarbonate levels of 5 mmol/L increased the odds of admission/observation 1.8-fold (95% CI: 1.1-3.0-fold).
In total, URVs occurred for 4.9% of the study population. URVs were correlated with the SEP results in the following way: 17.9% (95% CI: 6.1-36.9) of patients with either a low glucose level (<60 mg/dL) or a high potassium level (>5 mmol/L) had an URV, compared with 2.6% (95% CI: 0.7-6.6) of those with neither a low glucose level nor a high potassium level. A low glucose level (<60 mg/dL) alone was found for 17% of patients with URVs, compared with 3% of patients with URVs having glucose levels of ≥60 mg/dL.
Children presenting with acute gastroenteritis and resultant dehydration represent a common occurrence in the PED setting. The use of IVFs for rehydration usually occurs when a trial of oral rehydration has failed or the child presents with an altered level of consciousness. Rapid IV and nasogastric rehydration has been shown to be an effective way of treating these patients in an outpatient setting.5–8 There are limited studies supporting the usefulness of laboratory studies in evaluations of children with dehydration.9–11 A predictive marker of dehydration, indicating patients who could be treated with rapid IV rehydration and sent home, compared with those who should be admitted, could help expedite patient care. A previous study suggested that a serum bicarbonate concentration of ≤13 mmol/L identified most children who could not tolerate orally administered fluids after rapid IV rehydration and were hospitalized.4 Other studies have shown mixed results regarding the ability of SEP results to correlate with the degree of dehydration or to identify the children who require hospitalization.9–13 Our study examined the usefulness of the SEP in the treatment of children who presented with dehydration, all of whom received rapid IVF rehydration.
Frequency of SEP Abnormalities
Of the children in our study, 48% had ≥1 SEP abnormality. This large number included many values that fell outside the normal range but appeared to have little clinical significance. Most pediatric patients who present with dehydration exhibit isonatremia.14,15 Similarly, 96% of our study patients had isonatremic dehydration. Hypernatremia occurred for 3.3%, with 1 patient in our study having a serum sodium level of 168 mmol/L, which represented the highest value seen. This atypical syndromic patient was initially thought and then confirmed to have abnormal SEP results. The remaining hypernatremic patients had serum sodium levels of 150 to 155 mmol/L.
The most common SEP abnormality in our study was a low serum bicarbonate level. This finding is consistent with a prior study that also considered SEP abnormalities among dehydrated pediatric patients.16 In addition, our study demonstrated that low serum bicarbonate levels were correlated with younger age, particularly age of <1 year. This finding held true despite correction for the degree of dehydration and episodes of diarrhea.
Hypoglycemia was present for 9.9% of our patients, which is slightly greater than the 4.5% occurrence noted in one of the few studies that examined dehydration and hypoglycemia.17 This represented one reason why clinicians intervened or changed their management on the basis of SEP results.
How SEP Abnormalities Changed Medical Management
In 1996, the Provisional Committee on Quality Improvement, Subcommittee on Acute Gastroenteritis, of the American Academy of Pediatrics18 recommended that, because most episodes of dehydration are caused by diarrhea and are isonatremic, a SEP is unnecessary. It was suggested, however, that a SEP be considered for 1) children with moderate dehydration whose histories or physical findings were inconsistent with straightforward diarrheal episodes, 2) all severely dehydrated children, and 3) children receiving IVFs. Our findings indicate that SEP results altered clinical treatment for 10% of the children receiving IVFs because of dehydration, which supports this guideline.
Changes in the medical treatment of dehydrated patients can include immediate changes in therapy, consideration of other diagnoses, and secondary changes in disposition outcomes. The immediate medical management changes in our study included administration of glucose and potassium. There were 6 patients with glucose levels of <60 mmol/L (mean age: 18.4 months) who received IV glucose immediately and 2 patients who received potassium. Less-immediate management changes included prolonged therapy for patients with hypernatremia and the additional evaluation of a child with a low serum bicarbonate level and a positive history of previous IVF rehydration because of an inborn error of metabolism. Of patients with hypernatremia, 4 of 6 patients were either discharged after initial IVF rehydration or placed in the observation unit, given additional IVFs, and then discharged from the hospital successfully. Current treatment recommendations for IVF rehydration in hypernatremic dehydration suggest that patients require slow lowering of the serum sodium level (10–15 mEq/L per 24 hours), to avoid the development of brain edema.19 Four children with hypernatremia were rehydrated and discharged from the hospital from the PED because of clinical improvement, without additional monitoring of serum sodium levels. No patients with hypernatremia had an URV. Although this finding may suggest that not all patients with hypernatremia need the prolonged treatment and monitoring suggested in the current treatment recommendations, there were insufficient patients to allow any conclusions without additional study. In our study, SEP results identified some children with hypernatremic dehydration, and 67% of those patients did receive extended medical care.
Low serum bicarbonate levels also changed the medical management for some patients. The patients received prolonged IVFs and had a change in their disposition, ie, extended care.
Physicians' Decisions to Order a SEP and Sensitivity in Detecting a CSEA
Of the children in our study, 48% had ≥1 SEP abnormality, with 10.4% experiencing a change in treatment on the basis of the SEP results. In a multicenter prospective study, Rothrock et al16 attempted to establish criteria for identifying children who present to a PED with gastroenteritis/dehydration and a CSEA. As in our study, physicians were asked to predict which children would have a CSEA before they reviewed the results of laboratory studies. Sixty-eight percent of the children had a SEP abnormality, with 25% having a CSEA. As in our study, low serum bicarbonate levels were the most common finding. There was only moderate correlation between the physicians' suspicion that a CSEA was present and the actual presence of a CSEA, which indicates that clinical judgment alone was a poor predictor of the presence of a CSEA. There was similar poor sensitivity in the physicians' ability to predict which children would have a CSEA in our study (Table 4).
The usefulness of using the serum bicarbonate level as a predictive marker for medical management/disposition outcome has been reported. In a study of children with dehydration secondary to gastroenteritis, Reid and Bonadio4 showed that a serum bicarbonate level of ≤13 mmol/L was predictive of the need for hospitalization. Decisions regarding patient disposition were made 2 hours into the PED course. In contrast to those results, our study did not show that a low serum bicarbonate level was a marker for patient hospitalization. Nor did a serum bicarbonate level of >13 mEq/L indicate which patients could successfully tolerate orally administered fluids in the PED. Our study setting allowed for a longer period of PED treatment and observation (>2 hours), with continued IVF administration. We think that this type of management may be more reflective of many current ED settings.
Patients in our study were placed in the observation unit, with continued IVF administration, largely on the basis of clinical status (eg, an inability to tolerate orally administered fluids or continued significant fluid losses). Serum bicarbonate levels of ≤13 mmol/L were more prevalent in this group and represented a marker for children who received longer care and IVF rehydration. Of patients with serum bicarbonate levels of ≤13 mmol/L who were placed in the observation unit, 2 were admitted and the remaining 11 went home without an URV. For only 10% of all observed patients (5 of 48 patients) was the initial low bicarbonate level cited as the reason for placement in the observation unit by the attending physician. The majority of observation patients were placed in the observation unit because of a clinically assessed need for ongoing care, an inability to tolerate orally administered fluids, and the need for additional IVFs. The fact that there was not an increased number of URVs for these observation groups indicates the success of this approach for treating these dehydrated patients.
Children <1 year of age had lower serum bicarbonate levels, despite similar degrees of clinically assessed dehydration. Children with lower serum bicarbonate levels were also more likely to receive observation unit placement and additional IVFs. This age group (<1 year) may thus benefit from routine SEP monitoring. The use of the observation unit might have been partly attributable to greater concern for these younger patients, although it was not stated.
URVs occurred for a low percentage of our study population (4.9%). However, children with low glucose levels (<60 mmol/L) were more likely to have an URV (17% vs 3%). These children may require more conservative medical management.
Our PED includes a 6-bed observation unit, with annual visits by 2100 patients and an average length of stay of 10 hours. Dehydrated patients represent 21% of all patients. In this study, 27% of the enrolled children (48 children) were placed in the observation unit. The rate of URVs (6.3%, 3 of 48 patients) among patients who were placed in the observation unit and subsequently given additional IVFs was nonsignificantly (P = .148) higher (by 4.5%; 95% CI: −2.9 to 11.8) than that among patients who were discharged from the PED without observation (1.7%, 2 of 117 patients). This low rate of URVs supports the use of this approach. Patients placed in the observation unit received more IVFs, with a mean of 103 mL/kg, compared with 42 mL/kg for patients treated in the ED and then discharged from the PED (P < .0001). The use of our observation unit might have yielded a lower rate of admission, compared with other studies,4 but might be more reflective of current EDs that have observation capabilities beyond the 2-hour time period cited in that study.
Limitations to our study include the inability to comment significantly on patients with severe dehydration (≥10%), because they represented just 9% of our enrolled patients. However, our study results can be generalized to most ED settings, because the vast majority of patients have mild or moderate dehydration. There were insufficient numbers of cases of hyponatremia or hypernatremia to allow any conclusions regarding these subsets of dehydrated patients. Some attending physicians made disposition decisions on the basis of low bicarbonate levels only. In particular, some children were placed in the observation unit primarily on the basis of this laboratory value. It was thought that blinding the attending physicians to the SEP results represented an unethical study design. One hundred eighty-two patients were treated by a total of 8 attending physicians. Interobserver variability in the assigned degree of dehydration likely existed. In addition, some physicians might have treated more patients than others. Therefore, a certain amount of clustering might have occurred, which should be taken into account in interpretation of the results. Because clustering can increase CIs, the actual CIs might be slightly larger than reported. Telephone follow-up information was not obtained for 20% of the study patients; however, medical records for all patients were reviewed for repeat visits within our medical system. Our study population had a cutoff age of 9 years, because this represented the majority of dehydrated patients we encountered. The usefulness of the SEP among older children presenting with dehydration was not analyzed.
In addition, as the study progressed, some of the attending physicians indicated a new bias against the usefulness of ordering a SEP. This shift of opinion was expressed in the middle of the study, as we collected laboratory values for all patients, which was not the usual practice of all attending physicians at our institution.
The results of this study are institution specific. The use of an observation unit can decrease the number of inpatient admissions attributable to dehydration. This result can translate into financial savings. However, we did not attempt to analyze cost differences and reimbursement issues when comparing admissions and observation unit use. Our results are not applicable to ED settings without observation units or capability. It is thought, however, that the presence of observation units or the ability to observe patients for several hours within the ED itself is common in larger ED settings, and the results of this study could be generalized to those institutions.
In this study, children 2 months to 9 years of age who received IVFs for rehydration commonly had SEP abnormalities; 1 of 10 had an abnormality that affected clinical management. Low serum bicarbonate levels represented the most common abnormality and were more common among younger children. Attending physicians showed poor sensitivity in detecting dehydrated children with a CSEA. The routine ordering of a SEP for dehydrated children receiving IVFs not only would detect some apparently occult CSEAs but also would affect clinical management.
Correlation of SEP abnormalities with patient treatment and disposition suggests that additional time for IVF rehydration might be indicated for children with low serum bicarbonate levels. Children <1 year of age had lower serum bicarbonate values than did older children, despite having similar assessed degrees of dehydration. Patients with hypoglycemia at initial presentation who were then discharged from the hospital had a higher rate of URVs, which may suggest the need for prolonged therapy for these children and warrants additional study. In addition, clinicians could not tell, on the basis of clinical findings, which children were hypoglycemic. The use of an observation unit provided effective care for a subset of dehydrated patients, avoiding the need for hospital admission. The SEP has usefulness in the treatment of children receiving IVFs because of dehydration.
We thank Karen Leduc, RN, and the Nursing Research Committee at the Denver Children's Hospital for support with this project.
- ↵United States Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Division of Data Services. 1999–2000 National Ambulatory Medical Care Survey Dataset. Hyattsville, MD: United States Department of Health and Human Services; 2000
- ↵Cicirello HG, Glass RI. Current concepts of the epidemiology of diarrheal diseases. Semin Pediatr Infect Dis.1994;5 :163– 167
- ↵Johnson K. The Harriet Lane Handbook. 13th ed. St. Louis, MO: Mosby; 1993:101
- ↵Nager AL, Wang VJ. Comparison of nasogastric and intravenous methods of rehydration in pediatric patients with acute dehydration. Pediatrics.2002;109 :566– 572
- Teach SJ, Yates EW, Feld LG. Laboratory predictors of fluid deficit in acutely dehydrated children. Clin Pediatr.1997;7 :395– 400
- ↵Narchi H. Serum bicarbonate and dehydration severity in gastroenteritis. Arch Dis Child.1998;78 :70– 71
- ↵Fleisher G, Ludwig S. The Textbook of Pediatric Emergency Medicine. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2000:197
- ↵Finberg L, Kravath R, Hellerstein S, eds. Water and Electrolytes in Pediatrics: Physiology, Pathophysiology, and Treatment. Philadelphia, PA: WB Saunders; 1993:87, 125
- ↵American Academy of Pediatrics, Provisional Committee on Quality Improvement, Subcommittee on Acute Gastroenteritis. Practice parameter: the management of acute gastroenteritis in young children. Pediatrics.1996;97 :424– 435
- ↵Fleisher G, Ludwig S. The Textbook of Pediatric Emergency Medicine. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2000:817
- ↵Behrman R, Kliegman R, Jensen H. Nelson Textbook of Pediatrics. 16th ed. Philadelphia, PA: WB Saunders; 2000:213
- ↵McMillan JA. Oski's Pediatrics: Principles and Practice. 3rd ed. Philadelphia, PA: JB Lippincott; 1999:69
- Copyright © 2004 by the American Academy of Pediatrics