Diabetic ketoacidosis (DKA) is one of the most severe acute complications of type 1 diabetes mellitus (T1DM), an autoimmune disease characterized by an absolute deficiency of insulin and resultant hyperglycaemia. Several studies have been conducted in adults [1, 2], and these studies showed that leukocyte counts can add valuable information to reflect the presence of hyperglycaemic crisis [2], but there are few comparative studies on children with different severities of type 1 DKA [3]. Therefore, this study aimed to retrospectively assess the clinical manifestations, serum hormone levels, and biochemical levels of children with moderate to severe type 1 DKA during the acute and recovery phases to lay the foundation for a better understanding of DKA.
Patients and methods
Patients
Informed parental consent, patient consent, and approval (ethical approval number: 2020R132-E01) from the Hospital Ethics Committee were obtained before initiating the study. This study recruited 70 children with new-onset T1DM and moderate to severe ketoacidosis from the Department of Endocrinology at Shanghai Children’s Hospital, from 2015 to 2020. All patients met the diagnostic criteria for T1DM with diabetes symptoms, blood glucose ≥ 11.1 mmol/L, and low insulin and C-peptide levels [4]. Participants had no family history of diabetes, goitre, or obesity. Their DKA was diagnosed according to the consensus guidelines proposed by ISPAD [4]. The severity of DKA was divided into mild: venous pH < 7.3 or serum bicarbonate < 15 mmol/L; moderate: pH < 7.2, serum bicarbonate < 10 mmol/L; and severe: pH < 7.1, serum bicarbonate < 5 mmol/L. Because mild DKA can be treated with subcutaneous insulin in the general ward, children with mild DKA were not included in the study. The moderate and severe DKA groups included 33 and 37 patients, respectively.
Clinical manifestations
We evaluated clinical manifestations in the two groups by comparing differences in sex, age, body mass index (BMI), severity of hyperglycaemia (blood glucose concentration > 33.3 mmol/L) [4], presence of euglycaemic DKA (blood glucose concentration < 11.1 mmol/L) [5], low T3 syndrome, hypercholesterolemia, and hypertriglyceridaemia. C-peptide and insulin analyses were performed randomly before intravenous insulin administration. After initial rehydration, intravenous insulin was administered at a rate of 0.05–0.1 U/kg/h according to the latest guidelines [4]. Insulin and liquid infusion were adjusted according to the degree of dehydration correction, blood gas analysis, and blood glucose levels. Finally, we compared the time for acidosis correction and insulin dosage required to treat DKA after acidosis correction.
Hormonal and biochemical analyses
Blood biochemistry and blood gas analyses were conducted, and the following parameters were recorded for all patients: blood white blood cell (WBC) count and levels of insulin, C-peptide, blood glucose, glycosylated haemoglobin (HbA1c), diabetes autoantibodies, electrolytes, blood calcium, vitamin D, free triiodothyronine (FT3) with the normal range from 3.58 to 6.92 pmol/L, free thyroxine (FT4) with the normal range from 9.6 to 14.5 pmol/L, triiodothyronine (TT3) with the normal range from 1.34 to 3.7 nmol/L, thyroxine (TT4) with the normal range from 64.3 to 158.7 nmol/L, thyroid-stimulating hormone (TSH) with the normal range from 0.9 to 4.0 µIU/ml, triglycerides (TGs) with the normal range from 0 to 1.7 mmol/L, total cholesterol (TC) with the normal range from 0 to 5.72 mmol/L, high-density lipoprotein (HDL) with the normal range from 0.9 to 1.55 mmol/L, and low-density lipoprotein (LDL) levels with the normal range from 0 to 3.12 mmol/L. C-peptide and insulin analyses were performed randomly before intravenous insulin administration. Electrolytes, blood calcium, blood glucose, TGs, TC, HDL, and LDL were evaluated using detection kits (Beckman Coulter, Brea, CA, USA) and measured with an automatic biochemical analyser (AU5800 analyser, Beckman Coulter, Brea, CA, USA). Thyroid hormone concentrations were evaluated using detection kits (Beckman Coulter, Brea, CA, USA) and measured using an automatic immunoluminescence analyser (Unicel DxI 800, Beckman Coulter, Brea, CA, USA). C-peptide, insulin, and vitamin D levels were evaluated using detection kits (Roche Diagnostics, Mannheim, Germany) and measured using an automatic electrochemiluminescence immunoassay analyser (Cobas e601 analyzer, Roche Diagnostics, Mannheim, Germany). HbA1c levels were evaluated using detection kits (Tosoh Co., Tokyo, Japan) and measured using an automatic HbA1c analyser (HLC-723GX analyser, Tosoh Co., Tokyo, Japan). Diabetes autoantibodies were evaluated using detection kits (Yhlo Biotech Co., Ltd., Shenzhen, China) and measured using an automatic immunoluminescence analyser (iFlash 3000, Yhlo Biotech Co., Ltd., Shenzhen, China).
Statistical analysis
SPSS 26.0 software, manufactured by International Business Machines Corporation, was used to analyse the data. Nonparametric data were analysed using the Mann–Whitney U test and are presented as the median (25th and 75th percentiles). Normally distributed data were analysed using the t test and are presented as the mean ± standard deviation. Spearman’s rank correlation coefficient was used for correlation regression analysis, and Fisher’s test was used to compare ratios. A P value < 0.05 was considered to indicate a significant difference, with significance indicated as follows: *P < 0.05 and **P < 0.01.