· DKA patients needed a significantly higher insulin dosage than non-DKA patients for 2 years after diagnosis, which implies that DKA reflects the severity of disease progression by the deleterious effects on insulin secretion.

Diabetic ketoacidosis (DKA) is a common clinical presentation in patients with newly diagnosed type 1 diabetes mellitus (T1DM). Approximately 15%–70% of patients in Europe and North America [1,2] and approximately 38%–49% of patients in Korea [3,4] have been observed to have DKA at the initial diagnosis of T1DM. Efforts to achieve early diagnosis and proper management to improve emergency care for DKA have been made over the years, but DKA still places considerable economic burdens and sometimes results in life-threatening events. For instance, the estimated rate of mortality resulting from hyperglycemic crises, including DKA, was reported to be 1.8% in Korea from 2004–2014 [5].

DKA results from severe insulinopenia and usually occurs under stressful conditions, inducing metabolic decompensation; however, DKA sometimes presents without a specific cause, especially at the initial diagnosis of T1DM. DKA usually indicates the severity of insulin depletion attributed to islet destruction. Severe insulin deficiency and simultaneous activation of counter-regulatory hormones such as catecholamines, glucagon, and growth hormones result in excessive blood glucose levels and hepatic ketogenesis. Osmotic diuresis depletes fluid and electrolytes and increases free fatty acids, consequently eliciting severe dehydration and acidemia, sometimes resulting in fatal conditions. In DKA pathophysiology, the consequence of metabolic decompensation has been negatively associated with various endocrine functions, including thyroid [6,7] and growth hormone [8] functions. Although Hu et al. [9] recently reported that patients with DKA and euthyroid sick syndrome were shown to have poor glycemic control during the initial days of treatment, the long-term prognosis related to endocrinologic derangement in DKA patients has not yet been well documented.

Most DKA patients fully recover without major complications if appropriate insulin, fluid, and electrolytes are provided. However, deleterious effects of DKA have been observed depending on the severity of the disease or risk factors like a younger age, new-onset diabetes, and greater duration of symptoms. Cameron et al. reported that DKA at T1DM diagnosis results in morphologic and functional brain changes, which are detrimental to neurocognition [10]. Moderate to severe DKA at the onset of T1DM was also found to be detrimental to residual β-cell function, which was accompanied by poor long-term metabolic regulation [11,12]. Moreover, DKA at diagnosis has been suggested to be a poor prognostic factor for long-term glycemic control, regardless of demographic and socioeconomic factors [11,13]. Despite reports indicating the clinical significance of DKA at initial diagnosis, a limited number of studies have been conducted in South Korea with epidemiologic and geographic data, and none have assessed the prognosis or clinical outcomes of DKA during the follow-up period [3,14,15].

In this study, we aimed to investigate the clinical and biochemical differences in newly diagnosed T1DM patients with or without DKA at presentation. As differential severity at initial presentation possibly results in a differential clinical course, we analyzed possible diabetic complications (e.g., dyslipidemia, disturbed thyroid function, poor height growth, etc.) between DKA and non-DKA groups. In addition, we compared glycated hemoglobin (HbA1c) levels, residual islet functions, and insulin demands between DKA and non-DKA patients at initial diagnosis, as these factors have previously been reported to be major clinical consequences of initial DKA at T1DM diagnosis.

DKA has been suggested to be a poor prognostic factor for glycemic control, as it exacerbates the depletion of pancreatic islet cells [17], worsens residual β-cell function [18], and increases insulin requirements [19] in the early phase of T1DM progression. With regard to the long-term prognosis, a recent long-term prospective cohort study also reported that DKA at initial diagnosis could predict poor glycemic control [20]. However, there is little evidence that can be adapted to a variety of populations, including Asian patients, because previous studies were conducted in limited areas (i.e., mostly Western countries) with poor representation of different populations. Hence, in this study, we analyzed newly diagnosed T1DM patients with DKA and observed negative effects on clinical outcomes in a similar manner but in a different population. Specifically, we found that insulin demands were significantly higher in DKA patients than non-DKA patients for the first 2 years of follow-up after diagnosis in a South Korean population. These findings together support the notion that DKA at initial T1DM diagnosis exacerbates metabolic decompensation and decreases residual β-cell function [17,18], which could result in higher insulin requirements in the long term, eventually adversely influencing glycemic control and increasing the economic burden.

We did not find any effect of DKA at diagnosis on the glycemic control of T1DM patients. According to our data, a difference in glycemic control between DKA and non-DKA patients was not observed during the 5-year follow-up period; even at 3 months after T1DM diagnosis, HbA1c (%) did not vary significantly between the groups (data not shown). These findings conflict with those of a previous study by Duca et al. [20] where DKA as an initial presentation of T1DM was associated with a poor prognosis of glycemic control. One interesting difference between our study and that by Duca et al. [20] was initial HbA1c levels: in our investigation, initial HbA1c levels were relatively higher in both the DKA and non-DKA groups (12.50%±2.07% for non-DKA and 12.53%±1.96% for DKA) than in the study by Duca et al. [20] (9.9%±1.5% for DKA and 8.5%±1.4% for non-DKA). These discrepancies imply that the effects of DKA on long-term prognosis are not only possibly due to ethnical differences but also different physiological conditions at the time of initial T1DM diagnosis. In fact, a recent report suggested that the co-occurrence of various medical conditions with DKA, such as lactic acidosis, also result in worse clinical outcomes [21]. Although we speculated in this study that high HbA1c levels might suggest deleterious metabolic conditions to the residual islet cell function, their exact effects were not evaluated, especially in various ethnicities. Therefore, further studies will be needed to discern whether differential physiological conditions at diagnosis or ethnicity differentially affect the long-term prognosis.

Approximately 40% of newly diagnosed T1DM patients had DKA in our study, similar to the findings of other recent reports [3,4]. DKA patients more frequently presented with gastrointestinal symptoms such as hyporexia, abdominal pain, and nausea/vomiting; respiratory distress; and impaired consciousness. In addition, higher levels of serum glucose, total cholesterol, and LDL-c but lower levels of serum C-peptide, 24-hour urine C-peptide, and HDL-c were observed in patients with DKA compared to non-DKA patients. All symptoms are likely involved in the compensatory mechanisms of acid-base imbalance and/or the decompensation of metabolic crisis. On the other hand, compromised lipid profiles seemed to be associated with lipolysis of DKA. Together, these results suggested that DKA patients had more severe insulinopenia, which was also supported by decreased serum and urine C-peptides.

One of the interesting findings in our study was the endocrinological effect of DKA on T1DM patients. We found significantly low T3, free T4, and TSH levels in DKA patients, who also presented lower z-scores of serum IGF-1 and IGF-BP3 levels than non-DKA patients. Despite their low IGF-1 and IGF-BP3 levels at diagnosis, a difference in the z-score of height between DKA and non-DKA patients was not found (Fig. 2C). Possible mechanisms for the change in IGF-1/BP3 in T1DM have been suggested to include a reduction in hepatic growth hormone receptor or binding proteins in acute insulin deficiency, in which the decrease in IGF-1/BP3 was dependent on the severity of insulin deficiency [8]. It was also reported that IGF-1/BP-3 level may reflect improvements in metabolic control regardless of changes in growth hormone level [22,23]. Consistent with these previous reports, our present data showing an initial decrease in IGF-1/BP3 in patients with DKA may support the notion that IGF-1/BP3 reflects the severity of insulin deficiency, and the trend in the following data may also predict long-term glycemic control and prognosis. Future studies in larger cohorts will be needed to clarify the role these factors play in this matter.

The main limitation of this study is the conduct of a retrospective analysis based on data from a single center. Some records contained incomplete data regarding follow-up time or duration, which resulted in a large standard error of the mean in the linear mixed-model analysis. For instance, insulin demands seemed to be higher in DKA patients at all time points, but statistical significance was observed only in the first 2 years of follow-up. Future studies with a larger number of patients will be expected to resolve the limitations of incomplete data. Second, we postulated residual β-cell function with increasing administered insulin dosage; however, this parameter provided an indirect measurement of insulin secretion. Therefore, serial C-peptide measurements should be more appropriate for assessing the exact residual function of β-cells. Third, we did not adjust the analysis for the administration of different insulin analogs or the recurrence of DKA. Although recent advances in new insulin analogs and beneficial effects have been reported, we did not reflect their possible differential effects. In addition, the recurrence of DKA and its deleterious effect on residual insulin functions were not analyzed. In future studies, an analysis of the recurrence of DKA and residual β-cell function would also be helpful for better understanding the deleterious effects of DKA.

In summary, our analyses of serum blood glucose, lipid profile, serum/urine C-peptide, and endocrinological parameters including a thyroid function test and IGF-1/BP3 levels together supported that patients with DKA at diagnosis can be considered to have more severe insulin deficiency than those without DKA. Although the effect of DKA at diagnosis on the long-term prognosis or glycemic control was not made clear in our investigation, the negative effects of DKA on insulin demands or β-cell function were significant, which cannot be ignored in terms of clinical outcome. DKA, which can be life-threatening, is a major acute complication requiring early diagnosis and treatment, and it is associated with high socioeconomic costs in patients with T1DM. Therefore, the occurrence of DKA in children with T1DM should be closely examined, and these individuals should be supported with appropriate care to prevent recurrence.

This study was approved by the Institutional Review Board of Ajou University Hospital (AJIRB-MED-MDB-20-594), which waived the requirement for written informed consent due to the retrospective study design.