Factors associated with prolonged intensive care unit treatment and organ failure in pediatric patients with diabetic ketoacidosis

Jeffrey Eli Whang; Yi-Lei Wu; Jia-Yuh Chen; Chew-Teng Kor; Ming-Sheng Lee

doi:10.4103/prcm.prcm_25_22

ORIGINAL ARTICLE

Year : 2023 | Volume : 7 | Issue : 1 | Page : 13-19

Factors associated with prolonged intensive care unit treatment and organ failure in pediatric patients with diabetic ketoacidosis

Jeffrey Eli Whang¹, Yi-Lei Wu², Jia-Yuh Chen³, Chew-Teng Kor⁴, Ming-Sheng Lee⁵
¹ Department of Pediatric Pulmonology and Critical Care, Changhua Christian Children’s Hospital, Changhua, Taiwan
² Department of Pediatrics, Changhua Christian Children’s Hospital, Changhua, Taiwan
³ Department of Pediatrics, Changhua Christian Children’s Hospital, Changhua, Taiwan; Department of Post-Baccalaureate Medicine, College of Medicine, National Chung Hsing University, Taichung, Taiwan
⁴ Big Data Center, Changhua Christian Hospital, Changhua, Taiwan; Graduate Institute of Statistics and Information Science, National Changhua University of Education, Changhua, Taiwan
⁵ Department of Pediatric Pulmonology and Critical Care, Changhua Christian Children’s Hospital, Changhua, Taiwan; School of Medicine, Chung Shan Medical University, Changhua, Taiwan

Date of Submission	29-Dec-2022
Date of Decision	27-Jan-2023
Date of Acceptance	05-Feb-2023
Date of Web Publication	07-Apr-2023

Correspondence Address:
Dr. Ming-Sheng Lee
Department of Pediatric Pulmonology and Critical Care, Changhua Christian Children’s Hospital, No. 320, Xuguang Rd., Changhua 50010, Changhua
Taiwan

Source of Support: None, Conflict of Interest: None

Check

DOI: 10.4103/prcm.prcm_25_22

Abstract

Context: Patients with diabetic ketoacidosis (DKA) have potential complications, such as respiratory failure, cerebral edema, or acute renal injury, all of which can lead to a prolonged hospital course. Aims: This study identified risk factors for prolonged intensive care unit (ICU) stay and organ failure in pediatric patients with DKA. Materials and Methods: Patients with DKA aged <19 years admitted to the pediatric ICU of our hospital between June 2011 and May 2021 were enrolled. Demographic characteristics, initial Glasgow Coma Scale score, source of admission, biochemical values, ICU length of stay (LOS), and hospital LOS were collected. The primary outcome was to identify factors associated with prolonged (≥48 h) ICU treatment. The secondary outcomes were to identify factors associated with respiratory failure, cerebral injury, or acute renal failure. Results: This study enrolled 137 patients. Nonemergency room admission was associated with longer ICU LOS [adjusted odds ratio (aOR), 3.14; 95% confidence interval (CI) 1.01–9.82] compared with admission from the emergency room. Older age (aOR, 0.89; 95% CI, 0.80–0.99) and underweight (aOR, 0.33; 95% CI, 0.12–0.95) were associated with shorter ICU LOS. Conclusions: Recognizing the risk factors associated with prolonged ICU LOS in pediatric patients with DKA may help clinicians with the early identification of critical DKA cases.

Keywords: Diabetic ketoacidosis, intensive care unit, pediatric, length of stay

How to cite this article:
Whang JE, Wu YL, Chen JY, Kor CT, Lee MS. Factors associated with prolonged intensive care unit treatment and organ failure in pediatric patients with diabetic ketoacidosis. Pediatr Respirol Crit Care Med 2023;7:13-9

How to cite this URL:
Whang JE, Wu YL, Chen JY, Kor CT, Lee MS. Factors associated with prolonged intensive care unit treatment and organ failure in pediatric patients with diabetic ketoacidosis. Pediatr Respirol Crit Care Med [serial online] 2023 [cited 2023 May 29];7:13-9. Available from: https://www.prccm.org/text.asp?2023/7/1/13/373840

Key messages:

nonemergency room admission is associated with longer icu length of stay, whereas older age is associated with shorter icu length of stay in pediatric dka.

Introduction

A previous study using Taiwan’s National Health Insurance Research Database revealed that the diabetes mellitus (DM) prevalence rate in Taiwan from 2000 to 2009 in people younger than 19 years was approximately 0.06%–0.08%.^[1] Death in patients with insulin-dependent DM is predominantly due to diabetic ketoacidosis (DKA).^[2] Prior to the therapeutic use of insulin, the mortality rate of DKA was close to 100%; at present, it has decreased to 0.15%–0.3%.^[2],[3],[4] Patients with DKA having complications of cerebral edema, sepsis, shock, or acute kidney injury (AKI) can have a prolonged hospital stay and higher mortality.^[5] The median length of stay (LOS) in the hospital in pediatric patients with DKA is 2 days in the United States.^[6] However, the treatment of pediatric patients with DKA is different between Taiwan and other countries. A longer hospital course is usually observed due to more affordable medical expenses under Taiwan’s National Health Insurance program. Most pediatric patients with DKA are admitted to the intensive care unit (ICU) for close monitoring of their possible complications. Once DKA is resolved, thorough education about insulin injection, daily diet arrangement, and long-term follow-up plans are conducted at the ward. In this study, we identified the factors associated with prolonged ICU treatment and organ failure in pediatric patients with DKA.

Subjects and Methods

Study design

This is a retrospective cohort study. This study included all patients aged less than 19 years admitted to our hospital due to DKA between June 2011 and May 2021. A study based on the adult population indicated that the average LOS in the ICU was 2 days.^[7] Another pediatric population-based study concluded that prolonged metabolic acidosis, if not corrected within 24 h, in children with DKA led to a mean ICU LOS of 45 h.^[8] On the bases of these findings, we used 48 h as a reference to evaluate prolonged ICU LOS. The primary aim of this study was to identify the factors associated with ICU stay for >48 h. The secondary aim was to identify factors associated with respiratory failure, cerebral injury, and AKI. This study was approved by the institutional review board of our hospital, and the requirement of informed consent was waived.

Data collection

Medical records were collected by searching our hospital’s medical record database using ICD 9 (ICD 9 250.1 diabetes with ketoacidosis) and ICD 10 (E10.1 type 1 DM with ketoacidosis, E11.1 type 2 DM with ketoacidosis, E12.1 malnutrition-related DM with ketoacidosis, E13.1 other specified DM with ketoacidosis, E14.1 unspecified DM with ketoacidosis) codes. Patients whose laboratory data were missing or did not meet DKA criteria (serum glucose >200 mg/dL, venous pH <7.3 or bicarbonate <15 mmol/L, and presence of ketonemia or ketonuria) were excluded.

The following data were obtained from the medical records of each child: demographic characteristics, initial Glasgow Coma Scale (GCS) score, source of admission, initial biochemical value and creatinine changes, ICU LOS, and hospital LOS.

The patients were classified into three body types based on their body mass index (BMI) using the weight and height collected upon admission. BMI percentile was calculated using the BMI calculator for children and teens (https://www.cdc.gov/healthyweight/bmi/calculator.html). Because the BMI percentile is not recommended for children younger than 2 years of age according to the World Health Organization (WHO), they were evaluated using weight-for-length instead. The patients with BMI or weight-for-length below the 5th percentile were classified as being underweight. The patients with BMI higher than the 85th percentile or weight-for-length more than the 95th percentile were classified as being overweight. Those with intermediate values for these two variables were considered as healthy. Data published by the WHO were used as a reference (https://www.who.int/tools/child-growth-standards/standards/weight-for-length-height).

DKA severity was divided into mild, moderate, and severe according to the initial laboratory results as follows: mild DKA, pH < 7.3 or bicarbonate of 10–15 mmol/L; moderate DKA, pH <7.2 or bicarbonate of 5–10 mmol/L; and severe DKA, pH <7.1 or bicarbonate <5 mmol/L.

Serum sodium was corrected using the following formula: corrected serum sodium concentration = measured serum sodium concentration + 1.6 × (serum glucose concentration (mg/dL) − 100)/100.

Patients with lipase >648 U/L or an increase of more than three times the upper limit were considered with elevated lipase.

The patients were also divided into two groups based on the source of admission: the emergency room and the nonemergency room groups. The following patient groups were included in the nonemergency room group: patients whose DKA diagnosis was delayed and who were admitted to the ward first and those diagnosed as having DKA at other facilities but were unable to receive treatment immediately or required referral.

The ICU LOS of a patient with DKA depended on the time taken to resolve DKA (total CO₂ > 15 mEq/L; pH > 7.30 or sodium stable between 135 and 145 mEq/L). Once DKA was corrected, the patient could start oral intake, and their insulin therapy could be shifted from continuous intravenous infusion to subcutaneous injection; they would then be transferred to the ward.

Regarding secondary outcome data collection, patients who received mechanical ventilation support were considered as having respiratory failure. Those who presented with altered consciousness or signs of increased intracranial pressure received mannitol or glycerol as well as underwent brain imaging if cerebral edema was suspected. Any patient with a creatinine increase of 1.5-fold from the baseline or with decreased urine output, requiring diuretic use or renal replacement therapy, was regarded as having AKI.

Statistical analyses

Median values and interquartile ranges were calculated for all continuous variables. Bivariate analysis was used to identify factors associated with prolonged length of ICU stay (≥48 h). Between-group comparisons of continuous and categorical variables were performed using binary logistic regression and the chi-square test, respectively. Two-tailed P < 0.05 was considered statistically significant. Factors statistically significant associated with the prolonged length of ICU stay were further analyzed using multivariate logistic regression to calculate the adjusted odds ratio (aOR). All analyses were performed using SPSS v22 (IBM, Armonk, New York).

Results

Patient characteristics

We identified 150 admissions by using ICD 9 and ICD 10 codes in our hospital’s medical record database between June 2011 and May 2021. One patient was excluded due to incomplete medical records, and 12 patients were excluded because they did not meet the diagnostic criteria of DKA.

A total of 137 patients were included in this study. The median age (IQR) of the study population was 11 (8, 15) years. The study population comprised 50 boys and 87 girls. Approximately 56% (n = 77) of patients had new-onset diabetes and presented with DKA for the first time, whereas 46% (n = 60) patients were already diabetics on insulin therapy. Furthermore, 56% (n = 77) of the study population had healthy body weights, 31% (n = 43) were underweight, and 12% (n = 17) were overweight. As for DKA severity, 27% (n = 37) had mild DKA, 39% (n = 53) had moderate DKA, and 34% (n = 47) had severe DKA. In all, 86% (n = 118) of the patients were admitted from the pediatric emergency room, whereas the remaining were referred from other local hospitals or transferred from the ward. The median ICU LOS (IQR) of the study population was 43 (34, 50) h. The median hospital LOS (IQR) was 8 (6, 10) days.

Primary outcome

In this study, 92 patients (67%) had an ICU stay of <48 h, and 45 (33%) had an ICU stay of >48 h [Table 1]. No significant differences were observed between these groups in terms of sex, height, new diabetes diagnosis, DM type, initial white blood cell count, C-reactive protein elevation, serum osmolality, creatinine elevation, serum potassium level, and lipase elevation.

Table 1: Patient characteristics and demographic for ICU stay more than 48 h

Click here to view

Compared with the patients with ICU LOS <48 h, the patients with ICU LOS ≥48 h were significantly younger (P = 0.043); a lower proportion of these patients were underweight (P = 0.016); they had lower glycated hemoglobin (HbA1C, P = 0.025) and higher corrected serum sodium (P = 0.009) levels; and they had a lower proportion of emergency room admitted cases (P = 0.012). In addition, fewer severe DKA cases (P = 0.037) were found in this group, indicating that initial DKA severity was not associated with prolonged ICU stay.

In the multivariate logistic regression analysis, the only variable associated with prolonged ICU stay was admission from nonemergency room sources [aOR, 3.14; 95% confidence interval (CI) 1.01–9.82] [Figure 1]. Variables associated with shorter ICU stay were age (aOR, 0.89; 95% CI, 0.80–0.99) and being underweight (aOR, 0.33; 95% CI, 0.12–0.95). With every 1-year increase in age, the risk of ICU stay ≥48 h decreased by 11%.

Figure 1: Multivariate adjustment of factors associated with the likelihood of pediatric DKA patients staying in ICU for more than 48 h

Click here to view

Secondary outcomes

Five patients (3.6%) in this study had respiratory failure and were intubated during the ICU course. This group also had a significantly higher number of patients with newly diagnosed DM (P = 0.044) [Table 2]. Seven patients were suspected of having cerebral edema [Table 3], and five were suspected of having AKI. Factors associated with cerebral edema or AKI could not be identified. Although not statistically significant, patients with DKA who developed respiratory failure were suspected of cerebral edema, or had AKI were younger, and the majority of them were newly diagnosed as having DM.

Table 2: Characteristics and demographic of patients with acute respiratory failure

Click here to view

Table 3: Characteristics and demographic of patients suspected of cerebral edema

Click here to view

Discussion

In our study, patients admitted from nonemergency room sources exhibited a longer ICU stay, and older age and underweight were associated with a shorter ICU stay. These findings might be related to fluid status, because more severe dehydration at presentation was associated with a longer duration of insulin infusion.^[9] In our study, patients with ICU LOS ≥48 h had a higher serum sodium level than those with ICU LOS <48 h, indicating higher dehydration levels.

Patients who were admitted from the ward usually had delayed DKA diagnosis, whereas those referred from other facilities had a longer gap between DKA diagnosis and treatment. These conditions might worsen dehydration status, thereby necessitating a longer insulin infusion time. A study conducted in the adult DKA population revealed a similar result: a longer gap between admission and DKA consultation was associated with a longer hospital LOS.^[10] Notably, of the 19 patients from nonemergency room sources, 4 had a delayed diagnosis of DKA: they visited our emergency room but were not diagnosed with DKA and were admitted to the general ward first and then transferred to the pediatric ICU after DKA diagnosis. The remaining 15 patients were diagnosed as having DKA at other hospitals and were referred to our hospital for DKA management. All of the DKA cases admitted from ER were diagnosed during their first ER visit; none of them were diagnosed during the second visit. The delayed diagnosis rate for DKA was 3.2% in our hospital. However, these four patients did not stay in the pediatric ICU for more than 48 h. Therefore, the association of nonemergency room sources with pediatric ICU LOS for more than 48 h was mainly observed in referral patients.

Older children were able to access fluids to replenish volume losses, and underweight patients had a lower insensible water loss in terms of the surface area-to-volume ratio.^[11] The body surface area formula is the square root of [height (in cm) × weight (in kg)/3600]. Compared with a normal-weight patient, an underweight patient will have a lower body surface area. These measures reduced the likelihood of severe dehydration, resulting in a shorter ICU LOS. Moffett et al.^[12] found that underweight patients had shorter beta-hydroxybutyrate clearance times, which can cause faster DKA resolution, leading to a shorter ICU LOS.

Glucose level, admission HbA1C level, and initial DKA severity were not significant predictors of hospital LOS in the adult population in a previous study.^[10] A similar result was found in the pediatric population; HbA1C was not associated with hospital LOS, and no differences were observed in HbA1C levels between pediatric patients with DKA having newly diagnosed DM and those with known DM.^[13] In our study, the initial result indicated that lower HbA1C was associated with longer pediatric ICU LOS (≥48 h), but this association was not significant in the multivariate logistic regression.

In this study, the median ICU and hospital LOS were 43 h and 8 days, respectively. Everett et al.^[14] reported that the mean hospital LOS in pediatric patients with DKA in the United States was 2.38–2.51 days and was inversely related to medical expenses. The difference between their study and our study was likely due to the more affordable medical expenses under Taiwan’s National Health Insurance program.

Most pediatric patients with DKA have type 1 DM. Patients with type 1 DM account for 91% and 94% of the DKA admissions in our study and the aforementioned study in the United States, respectively.^[14] DKA is commonly seen in patients with newly diagnosed DM or in those with poor insulin compliance. The median age of newly diagnosed type 1 DM onset is approximately 10 years.^[15] Poor insulin compliance resulting in DKA admission is commonly found in the late teenage years.^[14] In our study, the majority (56%) of the patients with DKA in pediatric ICUs were newly diagnosed as having DM (median age: 11 years), which is consistent with the results of previous studies.^[15]

In our cohort, all five patients with respiratory failure also had unstable hemodynamics that required central line placement and inotropic agent support, and they were all newly diagnosed patients with DKA without a delayed diagnosis of DKA. A previous study listed the following risk factors for respiratory failure in patients with DKA: depletion of primarily intracellular ions (potassium, magnesium, and phosphate), pulmonary edema, respiratory tract infection, neuromuscular disease, and miscellaneous conditions.^[16] All patients with DKA and respiratory failure in our study had an electrolyte imbalance, including hypokalemia, hypocalcemia, hypermagnesemia, and hypophosphatemia. However, the small sample size precluded the identification of any of these conditions as risk factors.

According to the ISPAD Clinical Practice Consensus Guidelines 2018, the incidence of cerebral edema is 0.5%–0.9%, and the mortality rate is 21%–24% in patients with DKA.^[17] Moreover, cerebral edema is diagnosed when a patient fulfills one diagnostic criterion and two major criteria or one major and two minor criteria. The diagnostic criteria include an abnormal motor or verbal response to pain, decorticate or decerebrate posture, cranial nerve palsy, and abnormal neurogenic respiratory patterns (e.g., grunting, tachypnea, Cheyne–Stokes respiration, and apneusis). The major criteria include altered mentation or fluctuating level of consciousness, sustained heart rate deceleration (a decrease of more than 20 beats/min) not attributable to improved intravascular volume or sleep state, and age-inappropriate incontinence. The minor criteria include vomiting, headache, lethargy or not being easily arousable, diastolic blood pressure >90 mmHg, and age <5 years. These criteria have a sensitivity of 92% and a false-positive rate of only 4% for identifying cerebral edema. Cerebral edema was significantly associated with a lower initial partial pressure of arterial carbon dioxide and high initial serum urea nitrogen levels.^[18] The risk was higher in younger patients and those newly diagnosed as having DKA.^[19],[20] In our study, the incidence of suspected cerebral edema was 5% because we used broader criteria to identify patients with suspected cerebral injuries in an attempt to increase the case numbers. This led to an unusually high incidence rate, but no associated risk factors were found. However, all seven patients suspected of cerebral injury had hypocapnia (mean PaCO₂: 12. 8 ± 3.8 mmHg), and six had high initial serum urea nitrogen levels (mean: 37.2 ± 16.1 mg/dL).

The AKI incidence in our study (3.6%) was much lower than that in another cohort study (43%).^[21] Four of the five AKI cases in our study had unstable hemodynamics, but no associated risk factors could be identified. Myers et al.^[21] evaluated 1359 pediatric patients with admission for DKA, and they found that older age, higher initial blood urea nitrogen levels, higher heart rate, higher glucose-corrected sodium and glucose concentrations, and lower pH were associated with AKI. They defined AKI as an increase in serum creatinine by ≥0.3 mg/dL within 48 h, increase in serum creatinine to ≥1.5 times the baseline value within the last 7 days, or urine output of <0.5 mL/kg/h for 6 h. In addition, they used an estimated glomerular filtration rate of 120 mL/min/1.73 m² to calculate the expected baseline creatinine level using the Schwartz estimating equation. This method overcomes the fact that baseline creatinine values were not available for most episodes, which enabled the authors to identify more AKI cases. In our study, those patients who developed AKI had higher glucose-corrected sodium and higher serum glucose concentrations, but these patients were younger, and no difference in heart rate was identified.

Limitations

This was a retrospective study. The ICU LOS might be affected by the DKA treatment protocol at our hospital. Clinical courses and DM education programs are affected by Taiwan’s National Health Insurance program; thus, the results might have some differences from those in other countries. Some data, including DKA-associated body weight loss or DKA duration, could not be measured. Insufficient sample size precluded the identification of as risk factors for respiratory failure, cerebral injury, and AKI. Nevertheless, this study provided some insights into pediatric DKA management in Taiwan.

Conclusion

Nonemergency room admission for pediatric DKA patients was associated with prolonged pediatric ICU LOS. Early identification of DKA patients and early treatment initiation are vital for ensuring a short ICU LOS. Being older or underweight was associated with a shorter pediatric ICU LOS. Finally, patients with newly diagnosed DM presented as DKA were more likely to have respiratory failure, unstable hemodynamics, cerebral edema, and AKI.

Acknowledgement

This manuscript was edited by Wallace Academic Editing.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1.	Jiang YD, Chang CH, Tai TY, Chen JF, Chuang LM. Incidence and prevalence rates of diabetes mellitus in Taiwan: Analysis of the 2000–2009 Nationwide Health Insurance Database. J Formos Med Assoc 2012;111:599-604.
2.	Edge JA, Ford-Adams ME, Dunger DB. Causes of death in children with insulin dependent diabetes 1990–96. Arch Dis Child 1999;81:318-23.
3.	Curtis JR, To T, Muirhead S, Cummings E, Daneman D. Recent trends in hospitalization for diabetic ketoacidosis in Ontario children. Diabetes Care 2002;25:1591-6.
4.	Decourcey DD, Steil GM, Wypij D, Agus MS. Increasing use of hypertonic saline over mannitol in the treatment of symptomatic cerebral edema in pediatric diabetic ketoacidosis: An 11-year retrospective analysis of mortality. Pediatr Crit Care Med 2013;14:694-700.
5.	Poovazhagi V. Risk factors for mortality in children with diabetic ketoacidosis from developing countries. World J Diabetes 2014;5:932-8.
6.	Smaldone A, Honig J, Stone PW, Arons R, Weinger K. Characteristics of California children with single versus multiple diabetic ketoacidosis hospitalizations (1998–2000). Diabetes Care 2005;28:2082-4.
7.	Freire AX, Umpierrez GE, Afessa B, Latif KA, Bridges L, Kitabchi AE. Predictors of intensive care unit and hospital length of stay in diabetic ketoacidosis. J Crit Care 2002;17:207-11.
8.	Kimura D, Raszynski A, Totapally BR. Admission and treatment factors associated with the duration of acidosis in children with diabetic ketoacidosis. Pediatr Emerg Care 2012;28:1302-6.
9.	Ronsley R, Islam N, Ronsley C, Metzger DL, Panagiotopoulos C. Adherence to a pediatric diabetic ketoacidosis protocol in children presenting to a tertiary care hospital. Pediatr Diabetes 2018;19:333-8.
10.	Xu AC, Broome DT, Bena JF, Lansang MC. Predictors for adverse outcomes in diabetic ketoacidosis in a multihospital health system. Endocr Pract 2020;26:259-66.
11.	Somers MJ. In: Mattoo TK, Wilkie L, editors. Clinical Assessment and Diagnosis of Hypovolemia (Dehydration) in Children. Waltham, MA: UpToDate; 2020. Available from: https://www.uptodate.com/contents/clinical-assessment-of-hypovolemia-dehydration-in-children. [Last accessed on 12 Sep 2022].
12.	Moffett BS, Allen J, Khichi M, McCann-Crosby B. Impact of body habitus on the outcomes of pediatric patients with diabetic ketoacidosis. J Pediatr Pharmacol Ther 2021;26:194-9.
13.	Babiker A, Aljahdali GL, Alsaeed MK, Almunif AF, Mohamud MS, Al Mutair A, et al. Frequency and risk factors of diabetic ketoacidosis in a specialized children’s hospital, Riyadh: A cross-sectional study. Oman Med J 2022;37:e341.
14.	Everett EM, Copeland TP, Moin T, Wisk LE. National trends in pediatric admissions for diabetic ketoacidosis, 2006–2016. J Clin Endocrinol Metab 2021;106:2343-54.
15.	Jensen ET, Stafford JM, Saydah S, D'Agostino RB, Dolan LM, Lawrence JM, et al. Increase in prevalence of diabetic ketoacidosis at diagnosis among youth with type 1 diabetes: The SEARCH for diabetes in youth study. Diabetes Care 1573;202:1578.
16.	Konstantinov NK, Rohrscheib M, Agaba EI, Dorin RI, Murata GH, Tzamaloukas AH. Respiratory failure in diabetic ketoacidosis. World J Diabetes 2015;6:1009-23.
17.	Wolfsdorf JI, Glaser N, Agus M, Fritsch M, Hanas R, Rewers A, et al. ISPAD clinical practice consensus guidelines 2018: Diabetic ketoacidosis and the hyperglycemic hyperosmolar state. Pediatr Diabetes 2018;19:155-77.
18.	Glaser N, Barnett P, McCaslin I, Nelson D, Trainor J, Louie J, et al. Risk factors for cerebral edema in children with diabetic ketoacidosis. The Pediatric Emergency Medicine Collaborative Research Committee of the American Academy of Pediatrics. N Engl J Med 2001;344:264-9.
19.	Edge JA, Hawkins MM, Winter DL, Dunger DB. The risk and outcome of cerebral oedema developing during diabetic ketoacidosis. Arch Dis Child 2001;85:16-22.
20.	Rosenbloom AL. Intracerebral crises during treatment of diabetic ketoacidosis. Diabetes Care 1990;13:22-33.
21.	Myers SR, Glaser NS, Trainor JL, Nigrovic LE, Garro A, Tzimenatos L, et al. Frequency and risk factors of acute kidney injury during diabetic ketoacidosis in children and association with neurocognitive outcomes. JAMA Netw Open 2020;3:e2025481.