Practical Classification approach for newly diagnosed diabetes onset with ketosis or ketoacidosis in children and adolescent patients

doi:10.21203/rs.3.rs-4662137/v1

Download PDF

Research Article

Practical Classification approach for newly diagnosed diabetes onset with ketosis or ketoacidosis in children and adolescent patients

https://doi.org/10.21203/rs.3.rs-4662137/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Distinguishing diabetes diagnosis is fundamental to ensuring proper management of patients, but has been challenging, especially in newly diagnosed diabetes onset with ketosis or ketoacidosis.

Methods

A retrospective analysis was conducted on medical records from 2017/1/1 to 2020/4/30 among children and adolescents with new-onset diabetes accompanied with ketosis or ketoacidosis. Data was collected at diabetes onset and 2 years after discharge. Patients were classified as type 1 or 2 diabetes (T1DM; T2DM) based on the patient's medication and final diagnosis. The best diagnostic cut-off point was determined using receiver operating characteristic curves (ROCs) between T1DM and T2DM.

Results

Among 153 children and adolescents, 78 patients (51.0%) were diagnosed as T1DM and 75 patients (49.0%) were diagnosed as T2DM after 2 years of follow-up. There were significant differences in sex, age, family history, BMI, systolic and diastolic blood pressure, lipids, uric acid (UA), C-peptide, combined fatty liver ratio and any islet autoantibody-positive ratio at the time of onset (P < 0.05). In the ROC analysis, fatty liver, SBP, BMI, fasting/1-h/2-h C peptide at the time of onset performed well on diagnostic typing (ROC AUC = 0.79, 0.83, 0.92, 0.94, 0.96, and 0.95 respectively; Optimal cut point = 1.5, 110.5, 21.0, 0.5, 1.0 and 2.0).

Conclusions

This study provides a practical clinical approach to the diagnosis and classification of diabetes. Caution is needed in C-peptide, BMI, SBP and fatty liver at the time of onset, which have effective diagnostic values.

Diabetic ketosis (DK)

Diabetic ketoacidosis (DKA)

Clinical characteristics

diabetes typing

Adolescents

Distinguishing diabetes diagnosis is fundamental to ensuring proper management of patients, but has been challenging, especially in newly diagnosed diabetes onset with ketosis or ketoacidosis.
We retrospectively collected the clinical data from newly 153 diagnosed diabetic children and adolescents with DK/DKA accompanied at the onset and followed up for 2 years, and further figures out the effective diagnostic metrics in effort to shed light on better classify diabetes.
Our results indicate that C-peptide, BMI, SBP and fatty liver at the time of onset have effective diagnostic values. It provides a practical classification approach for newly diagnosed diabetes onset with ketosis or ketoacidosis in children and adolescents, by taking advantage of their characteristics at presentation, which are readily available in clinical practice.

Diabetes is one of the most common chronic diseases in childhood. According to the latest estimates, more than 1.2 million children and adolescents worldwide are living with diabetes, and the number of newly diagnosed cases each year is around 180,000(1). As guidelines recommended (2–4), the classification of diabetes in children and adolescences at diagnosis is typically based on their characteristics at presentation and genetic measurements for the specific types. However, increasingly, the ability to make a clinical diagnosis has been hampered due to the overlapping clinical characteristics between type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM) (5).

Diabetic ketosis (DK) and ketoacidosis (DKA) occur commonly at the onset of diabetes and have long been considered as a key clinical feature of T1DM, particularly among children and adolescents, with DKA frequencies of 15 ~ 70% (6, 7). While it had been demonstrated recently that obese adolescents with a clinical picture suggestive of T2DM can present in DK/DKA of varying degrees with the prevalence of 5 ~ 25% (8, 9). Of note, the near-normoglycemia remission has been reported in some patients and may last for months to years. This presentation and clinical course of diabetes has been called as diabetes type 1B, idiopathic T1DM, atypical diabetes, or ketosis-prone T2DM in recent years (10, 11). It has also been well described in Blacks and demonstrated to affect other populations at high risk of T2DM such as Chinese, Japanese and Hispanic(12–14).

The appropriate classification of diabetes is important for determining therapy, and thus, it is critical for clinicians differentiate between the types of diabetes as early as possible. Previously, a group of clinical characteristics, including c-peptide, body mass index (BMI) and onset age, had been proven to be effective diagnostic indicators for diabetes classification among newly diagnosed diabetic adults with DK/DKA accompanied, or among mixed cohorts of adults with the small proportion of youth included. However, it is important to note that the evidence in adults cannot be simply promoted to children and adolescents since there are vast differences between adult-onset and pediatric diabetes, including pathophysiology, development, and response to therapy. In addition, there is currently sporadic evidence to differentiate newly diagnosed diabetes in children and adolescent patients with DK/DKA accompanied, with even none in Asians (15). Therefore, in this study, we retrospectively collected the clinical data from newly diagnosed diabetic children and adolescents with DK/DKA accompanied at onset and followed up for 2 years, and further figures out the effective diagnostic metrics in an effort to shed light on better classifying diabetes.

This was a hospital-based, retrospective study with data collected from newly diagnosed diabetic children and adolescents with ketosis or ketoacidosis at the onset in Henan Provincial People’s Hospital. The medical records of all episodes of DK/DKA between January, 2017 and April, 2020 were reviewed. A flowchart of patient selection was presented in Fig. 1. This study was conducted according to the guidelines outlined in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Ethics Committee of the Henan Provincial People’s Hospital.

Participants

The inclusive criteria are as follows: 1) The onset age is under 18 years; 2) diagnosed with DKA, according to the ICD code of an electronic hospital system (EHRs), or with the biochemical criteria of DKA recommended by guidelines(2, 16): a. hyperglycemia (blood sample glucose > 11 mmol/L [≈ 200 mg/dL]), serum blood bicarbonate < 18 mmol/L or pH < 7.30, ketonemia (blood β-hydroxybutyrate ≥ 3mmol/L) or moderate or large ketonuria; those with hyperglycemia, ketonuria and ketonemia while without acidosis were also included due to the presence of DK; 3) having medical records at both the onset and after 2 years of diabetes. Those with other types of diabetes, including the sedentary diabetes induced by drugs or pancreatitis, neonatal diabetes and genetically induced diabetes, were excluded from this study.

Data collection

Data at the onset of diabetes was collected by endocrinologists from EHRs via Excel. Information regarding patient demographics, physical examinations (BMI; systolic and diastolic blood pressure [SBP; DBP]), medical histories related to diabetes [family history, precipitating factors for DK/DKA, the presence of fatty liver, treatment, and any acute or chronic complications] were collected. Biomedical characteristics, such as glucose levels at admission, values of glycated hemoglobin A1c [HbA1c], triglyceride [TG], total cholesterol [TC], serum uric acid [UA], serum creatine [sCr], and high/low-density lipoprotein cholesterols [HDL-C; LDL-C] were collected. The qualitative results of islet autoantibodies, including glutamic acid decarboxylase autoantibody (GADA), insulinoma-associated antigen-2 autoantibody (IA-2A), insulin autoantibody (IAA), zinc transporter-8 autoantibody (ZnT8A), and islet-cell antibody (ICA) were also collected. In addition, patients continued to receive intensive insulin therapy for 7 to 10 days (average 8.5 days) after correction of DK/DKA. Then, once the blood glucose level reached the standard and steadily, the C-peptide release test of oral glucose tolerance was conducted, and C-peptide at 0h, 0.5h, 1h, 2h and 3h were collected.

Follow-up and classification

Information on treatment and diagnosis after 2 years was followed up via the medical records which collected in the outpatient and inpatient EHRs. Patients were finally diagnosed with T1DM if they had a prior diagnosis of T1DM at the onset and continuously received insulin treatment during 2 years. Alternatively, patients were also classified as having T1DM if they had a repeated C peptide below 0.6 nmol/L and received insulin treatment during the 2-year follow-up. On the other hand, patients were finally assigned to T2DM group if they had a previous diagnosis of diabetes and were managed solely with diet and exercise, or if they were prescribed oral hypoglycemic agents (OHAs). Additionally, patients who failed to comply with their insulin regimen for more than 4 weeks without experiencing recurrent DK/DKA at any point during their illness were also assigned to the T2DM group.

Statistical analysis

Statistical analyses of the data were performed using SPSS 26.0 (Chicago, IL). Statistical significance was defined as p value < 0.05. Characteristics were presented as mean (SD) or median with interquartile range (IQR), given in parentheses for normalized continuous or skewed variables. Categories of variables were presented as counts (percentages). One-way analysis of variance (ANOVA) or the Mann–Whitney U-test was used for comparisons of quantitative variables among groups. Chi-squared test was performed to assess differences in proportions across groups. The receiver operator characteristics (ROC) curve and logistic analysis were generated to compare the diagnostic performance of the potential clinical characteristic, with the sensitivities, specificities and optimum cut-off points calculated for diabetes classification.

Clinical presentations and distributions of included patients

Table 1 summaries the demographic and clinical characteristics of patients with DK/DKA at the onset of diabetes enrolled in this study. A total of 221 patients with medical records of DK/DKA when newly diagnosed as diabetes were included in the analysis, with 153 (69.23%) patients meeting the inclusion criteria (Fig. 1). Among these 153 newly diagnosed diabetic patients with DK/DKA onset, the mean age was 12.65 ± 3.68 years, with females accounting for 43.7% (67/153) of the sample. The average BMI was 21.66 ± 5.69 kg/m², random glucose levels averaged 18.36 (15.00,22.85) mmol/L, and HbA1c levels were measured at 12.52 ± 2.38%. According to the diagnosis at discharge, 72 (47.1%) patients were diagnosed with T1DM, 46 (30.1%) patients were diagnosed with T2DM and the remaining 35 patients (22.8%) could not be “classified”.

Table 1

Clinical characteristics of newly diagnosed diabetic children and adolescents with ketosis/ketoacidosis onset.
	Total (N = 153)	T1DM (n = 72)	T2DM (n = 46)	Unclassified (n = 35)	F/χ2/H	P-value
Gender (M/F) (n)	86/67	32/40	34/12	20/15	9.92	0.007
Age at the onset (yrs)	12.65 ± 3.68	11.36 ± 3.66	14.41 ± 2.82	13.00 ± 3.36	11.17	< 0.001
Family history (n)	58(37.9)	17(23.6)	25(54.3)	16(45.7)	12.44	0.002
BMI (kg/m²)	21.66 ± 5.69	17.93 ± 3.98	27.30 ± 3.71	21.92 ± 4.58	75.26	< 0.001
SBP (mmHg)	114.00 ± 14.00	106.40 ± 11.48	125.26 ± 11.04	112.91 ± 13.38	35.81	< 0.001
DBP (mmHg)	72.00 ± 10.00	69.90 ± 10.55	74.37 ± 8.58	73.51 ± 9.94	3.37	0.037
Hyperglycemia on admission
Random Glucose (mmol/L)	18.36 (15.00, 22.85)	19.88(16.63, 25.37)	15.38(11.88,19.75)	18.36(15.91,24.00)	14.79	0.001
HbA1c (%)	12.52 ± 2.38	13.10 ± 2.42	11.60 ± 2.37	12.55 ± 1.85	6.06	0.003
Lipids on admission
TC (mmol/L)	4.41 (3.82, 5.09)	4.32(3.70,4.99)	4.58(4.03,5.29)	4.23(3.76,5.04)	3.51	0.173
TG (mmol/L)	1.35 (0.91, 2.15)	1.06(0.80,1.61)	1.85(1.28,3.29)	1.34(0.81,2.42)	23.02	< 0.001
HDL-C(mmol/L)	1.05 (0.86, 1.24)	1.10(0.88,1.37)	0.97(0.76, 1.09)	1.05(0.90,1.28)	11.84	0.003
LDL-C(mmol/L)	2.62 ± 0.80	2.44 ± 0.71	2.92 ± 0.84	2.60 ± 0.81	5.52	0.005
Serum UA (umol/L)	282.5(217.5, 422.3)	239.0(179.5,307.0)	420.5(330.8,604.5)	260.0(226.0,369.0)	36.20	< 0.001
Serum Creatine (umol/L)	37.00 (31.00, 49.00)	35.0(26.25,45.50)	42.00(34.75,53.00)	39.00(32.00, 46.00)	0.61	0.544
Beta-cell function
C peptide 0h (nmol/L)	0.49(0.14,1.35)	0.15(0.10,0.32)	1.55(1.10,2.38)	0.68(0.36,1.27)	79.95	< 0.001
C peptide 0.5h (nmol/L)	0.69(0.21,2.11)	0.21(0.10,0.33)	2.42(1.43,4.28)	1.01(0.52,1.97)	90.61	< 0.001
C peptide 1h (nmol/L)	0.97(0.30,3.15)	0.30(0.10,0.62)	3.96(2.45,6.55)	1.42(0.79,3.01)	92.28	< 0.001
C peptide 2h (nmol/L)	1.74(0.44,4.11)	0.44(0.15,0.95)	6.29(3.62,8.59)	1.84(1.12,3.42)	93.72	< 0.001
C peptide 3h (nmol/L)	1.77(0.54,4.26)	0.54(0.17,1.08)	5.10(3.40,7.02)	2.12(1.22,3.66)	94.30	< 0.001
Islet antibody ≥ 1 positive (%)	31(20.3)	26(36.1)	2(4.3)	3(8.6)	21.37	< 0.001
GADA positive (%)	30(19.6)	25(34.7)	2(4.4)	3(8.6)	19.60	< 0.001
IAA positive (%)	1(0.7)	1(1.4)	0(0.0)	0(0.0)	1.12	0.572
ICA positive (%)	7(4.6)	6(8.3)	0(0.0)	1(2.9)	4.50	0.096
ZnT8 positive (%)	3(2.0)	2(2.8)	0(0.0)	1(2.9)	1.29	0.525
IA2Ab positive (%)	8(5.2)	7(9.7)	0(0.0)	1(2.9)	5.78	0.056
Fatty Liver (%)	52(34.0)	5(6.9)	38(82.6)	9(25.7)	73.01	< 0.001
Precipitating factors (%)	117(76.5)	55(76.4)	37(80.4)	25(71.4)	0.90	0.639
Treatment at discharge
Insulin (%)	99(64.7)	68 (94.4)	9 (19.6)	22 (62.9)	91.69	< 0.001
Insulin + OHA (%)	17(11.1)	2 (2.8)	10 (21.7)	5 (14.3)
OHA (%)	21(13.7)	2 (2.8)	16 (34.8)	3 (8.6)
GLP-1(%)	10(6.5)	0 (0.0)	10 (21.7)	0 (0.0)
Drugs withdrawal (%)	6(3.9)	0 (0.0)	1 (2.2)	5 (14.3)
Abbreviations: T1DM: Type 1 diabetes mellitus; T2DM: Type 1 diabetes mellitus; BMI: Body mass index; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; HbA1c: Glycated hemoglobin A1c; TC: Total cholesterol (reference range: 3.10–5.70 mmol/L); TG: Triglyceride (reference range: 0.34–1.92 mmol/L); HDL-C: High-density lipoprotein cholesterol (reference range: 0.78-2.00 mmol/L); LDL-C: Low-density lipoprotein cholesterol (reference range: 2.07–3.10 mmol/L); UA: Urine acid (reference range: 90–420 umol/L); GADA: Glutamate decarboxylase antibody; IAA: Insulin autoantibody; ICA: Islet-cell antibody; ZnT8A: Zinc transporter 8 autoantibody; IA2Ab: Islet antigen-2 (IA-2) autoantibody; OHA: Oral-hypoglycemic agents; GLP-1: Glucagon-like peptide-1.

Clinical characteristics between groups after 2 years

After the 2-year follow-up, there were 113 patients maintaining their initial diagnosis, with 67 (93.06%) patients diagnosed with T1DM and 46 (100.0%) patients diagnosed with T2DM. For those diagnosed with T1DM at onset, 5 patients were subsequently diagnosed with T2DM, among which 2 patients were treated with insulin in combination with OHAs, 2 patients were treated with OHAs alone, and the other one discontinued any anti-hyperglycemic treatment. For those with “un-classified” diabetes at onset (N = 35), 11 patients were ultimately diagnosed with T1DM, with the majority of them receiving insulin treatment and only one patient receiving a combination of insulin and OHAs. The remaining 24 patients were finally diagnosed with T2DM, with 16 patients receiving OHAs, 1 receiving GLP-1 treatment, 3 patients receiving a combination of insulin and OHAs, 2 patients maintaining unmedicated, and 2 patients maintaining basic insulin treatment only.

The between-group clinical characteristics after 2 years were shown in Table 2. Among those who were finally diagnosed with T1DM (n = 78) and T2DM (n = 75), it was observed that, compared to T2DM group, glycemia level was significantly higher in T1DM group, with a random glucose level of 19.60 (16.48,24.93) vs. 16.83 (13.80,22.00) mmol/L (P = 0.005) and a higher HbA1c value of 13.45(11.88,14.63) vs. 12.10(10.10,13.50) % (P = 0.001). Additionally, the positive rates of any islet antibody were also higher in the T1DM group (33.3% vs. 6.7%, P < 0.001), particularly the positive rate of GADA with 32.1% vs. 6.8% (P < 0.001). While for those finally diagnosed with T2DM, male patients predominated (66.7% vs. 46.2%, P = 0.011), and there was a significantly higher age at the onset (14.00[13.00,16.00] vs. 12.00[8.00,14.00] years, P < 0.001), BMI (25.77 ± 4.20 vs. 17.71 ± 3.84 kg/m², P < 0.001), concomitant rate of fatty liver (64.0% vs. 5.1%, P < 0.001), percentage of family history of diabetes and blood pressure than T1DM group. In terms of β-cell function, T2DM had significantly higher C peptide levels in both fasting and within 3-h postprandial periods (P all < 0.05).

Table 2

Comparison of clinical characteristics between the two groups after follow-up.
	T1DM	T2DM	t/χ²/Z	P-value
N (%)	78(51.0)	75(49.0)
Gender (M/F)	36/42	50/25	6.536	0.011
Age at the onset (yrs)	12.00(8.00,14.00)	14.00(13.00,16.00)	-4.764	<0.001
Family history of diabetes (n/%)	16(20.5)	42(56.0)	20.457	<0.001
BMI (kg/m²)	17.71 ± 3.84	25.77 ± 4.20	-12.396	<0.001
SBP (mmHg)	105.81 ± 10.51	121.63 ± 13.19	-8.221	<0.001
DBP (mmHg)	69.71 ± 9.95	74.53 ± 9.52	-3.064	0.003
Random glucose (mmol/L)	19.60(16.48,24.93)	16.83(13.80,22.00)	-2.823	0.005
HbA1_C (%)	13.45(11.88,14.63)	12.10(10.10,13.50)	-3.202	0.001
TC (mmol/L)	4.25(3.68,4.85)	4.55(3.95,5.27)	-1.745	0.081
TG (mmol/L)	1.06(0.78,1.57)	1.78(1.17,2.85)	-4.654	<0.001
HDL-C (mmol/L)	1.07(0.88,1.37)	1.02(0.79,1.14)	-2.176	0.03
LDL-C (mmol/L)	2.45 ± 0.73	2.80 ± 0.82	-2.796	0.006
Serum UA (mmol/L)	237.00(180.5,303.25)	379.00(260.00,480.00)	-5.245	<0.001
Serum Creatine (mmol/L)	33.50(26.75,45.00)	42.00(34.0,52.00)	-3.591	<0.001
C peptide 0h (nmol/L)	0.15(0.10,0.33)	1.35(0.73,2.20)	-9.411	<0.001
C peptide 0.5h (nmol/L)	0.21(0.10,0.34)	2.11(1.27,4.04)	-9.975	<0.001
C peptide 1h (nmol/L)	0.32(0.10,0.64)	3.13(1.62,5.20)	-9.864	<0.001
C peptide 2h (nmol/L)	0.50(0.16,1.04)	3.95(2.43,7.33)	-9.574	<0.001
C peptide 3h (nmol/L)	0.54(0.17,1.16)	4.18(2.59,6.36)	-9.714	<0.001
Positive rate of any islet antibody (%)	26(33.3)	5(6.7)	16.829	<0.001
GADA positive (%)	25(32.1)	5(6.8)	15.337	<0.001
IAA positive (%)	1(1.3)	0(0.0)	0.955	0.328
ICA positive (%)	7(9.0)	0(0.0)	6.962	0.008
ZnT8 positive (%)	3(3.8)	0(0.0)	2.903	0.088
IA2Ab positive (%)	8(10.3)	0(0.0)	8.011	0.005
Fatty Liver (%)	4(5.1)	48(64.0)	59.066	<0.001
Precipitating factors (%)	58(74.4)	59(78.7)	0.394	0.530
Abbreviations: T1DM: Type 1 diabetes mellitus; T2DM: Type 1 diabetes mellitus; BMI: Body mass index; HbA1c: Glycated hemoglobin A1c; TC: Total cholesterol (reference range: 3.10–5.70 mmol/L); TG: Triglyceride (reference range: 0.34–1.92 mmol/L); HDL-C: High-density lipoprotein cholesterol (reference range: 0.78-2.00 mmol/L); LDL-C: Low-density lipoprotein cholesterol (reference range: 2.07–3.10 mmol/L); UA: Urine acid (reference range: 90–420 umol/L); GADA: Glutamate decarboxylase antibody; IAA: Insulin autoantibody; ICA: Islet-cell antibody; ZnT8A: Zinc transporter 8 autoantibody; IA2Ab: Islet antigen-2 (IA-2) autoantibody.

ROC curves and logistic regression analysis for classifying diabetes

The performance of each metric in classifying diabetes, as indicated by the estimated AUC and its 95%CI, is presented in Fig. 2. The β-cell function, as measured by the C peptide level within 3 hours, exhibited the highest AUC values for diabetes classification, ranging from 0.940 (95%CI: 0.903 ~ 0.976) ~ 0.961 (95%CI: 0.932 ~ 0.989) (P all < 0.05). The BMI, with an optimal cut-off point of 20.95 kg/m², demonstrated a significant prognostic value with an AUC value of 0.918 (95%CI: 0.874 ~ 0.962) and the sensitivity and specificity was 88.0% and 83.1% respectively. Both SBP and the presence of fatty liver demonstrated a significant and relatively high prognostic value, with a similar value of 0.826 (95%CI: 0.761 ~ 0.890) and 0.794 (95%CI: 0.720 ~ 0.869). The AUCs of the serum UA, age at the onset, and TG in diabetes classification were all found to be significantly moderate, with a gradual decrease in value from 0.743 (95%CI: 0.663 ~ 0.824) for UA, 0.719 (95%CI: 0.638 ~ 0.799) for age to 0.716 (95%CI: 0.635 ~ 0.797) for TG (all P < 0.05). Conversely, the other metrics, such as gender, family history of diabetes, presence of islet autoantibodies, DBP, TC, HDL-C/LDL-C, and creatine exhibited poor performance and did not demonstrate prognostic ability, with all AUCs being less than 0.7. Logistic regression analysis further confirmed that BMI (P = 0.003), SBP (P = 0.042) and C peptide levels at 0h (P = 0.046) and 0.5h (P = 0.001) were significantly effective in classifying diabetes.

This study is undertaken to investigate the clinical metrics that effectively shed light on classifying types of diabetes among newly diagnosed diabetic patients with DK/DKA onset. It had demonstrated that C peptide, BMI, SBP and the presence of fatty liver can differentiate diabetes types of newly diagnosed diabetes with DK/DKA onset among children and adolescences. These metrics may assist in the clinical diagnosis and classification of diabetes, particularly in cases with overlapping characteristics.

Insulin deficiency is clinically evident through deteriorating glucose control and increased susceptibility to DK or DKA. In this study, similar to previous studies conducted among both adults and those with younger ages (12, 14, 17), the levels of C-peptide remained the most effective value for distinguishing between T1DM and T2DM during the entire duration of the 3-hour tests. However, it is worth noting that the cut-off point for C-peptide in this study was higher than the other studies, with a fasting level of 0.4 nmol/L and 2-h level of 2.0 nmol/L, whereas other studies among adults reported a fasting C-peptide levels of around 0.2 nmol/L and the 2-h level of approximately 0.5 nmol/L (18). Previously, there was evidence indicating that C-peptide levels at diagnosis were elevated with increasing age within each BMI group, highlighting the independent influence of age on C peptide (19–21). Thus, the discrepancy observed in our study may be attributed to the variations in age, as we included the youngest age of patients among all the studies. Besides, it is worth noting that C-peptide levels may initially rise during the diagnosis of pediatric T1DM, leading to a remission period, but subsequently decline rapidly in the first years after diagnosis. Consequently, employing a relatively high cut-off point for C peptide may enhance the sensitivity and specificity of classification. It is of great importance for pediatrics to pay more attention to those with moderately reserved C peptide.

Overweight has been proved to be associated with accelerated progression to T1DM in children and adolescents, in addition to its well-known association with T2DM. In this study, similar to previous studies focusing on adults with ketone-prone diabetes (22), BMI was proved to have significantly effective value of the classification between T1DM and T2DM among children and adolescents with DK/DKA at the onset, with the ranks of ROC analysis just after C-peptide. Furthermore, unlike most other studies(13, 14), our study had further included the rate of fatty liver as an evaluation of vesical fat rather than solely relying on subcutaneous fat (estimated by BMI), and it has been observed that the rateincidence of fatty liver does have a relatively high effective value for diabetes classification, particularly in relation to diabetes types in the context of obesity. This highlights the importance of accurately distinguishing between different types of diabetes, especially in the presence of childhood obesity, as it contributes to an overall increase in the population and in youths with diabetes.

The autoimmune antibodies, one of the emblematic metrics of T1DM, had been reported to be positive among several youths with T2DM (23). For ketone-prone diabetes, since it was discovered, the positive status of antibodies has been considered a major value metric for diabetes classification. However, in our study, even though there was a statistically significant difference in the positive rate of islet antibodies between T1DM and T2DM, the relatively small sample size may have contributed to the less effective value in the ROC analysis. Furthermore, as presented by previous studies, the islet of antibodies can be also negative in up to 20% of otherwise classical T1DM, particularly in children. The reported lower sensitivity or false negative rate may also contribute to the results.

A strength of the present study is the relatively large number of participants in both the confirmation and prognosis cohort, which had largely reduced the bias from patient selection and post-hoc analysis between comparison of clinical characteristics at the onset had been also made in the present study. In addition, the clinical characteristics employed in our study were those commonly utilized in clinical practice, enhancing the practicality of the obtained findings. The major limitation of our study was the retrospective collection of data from medical records rather than the prospective ones. Therefore, several major metrics, including the nature of islet antibodies, usually missing after 2-year follow-up, which might potentially bias the results. Furthermore, the majority of patients do not receive genetic testing, potentially resulting in the misdiagnosis of monogenic diabetes and skewing the outcomes. Another limitation of this study is that it is an investigation of DK/DKA clinical profiles from a single institution in the central province of China. To validate the effectiveness of prognostic metrics for classification, it is necessary to conduct further investigations with larger cohorts and, most importantly, prospective followed up studies.

In patients with ambiguous clinical features in whom misclassification of diabetes type occurs at initial presentation, the time between presentation and correct classification poses significant potential risks. Therefore, this study provides a practical clinical approach to the diagnosis and classification of diabetes. Caution is needed in C-peptide, BMI, SBP and fatty liver at the time of onset, which have effective diagnostic values.

Author Contributions: Conceptualization, HXL and CSS; Methodology, HXL and YW; Data cu-ration, MW,BZ, CXM, LLC, QHW, and ZFC; Writing—original draft preparation, HXL and YW; writing—review and editing, HXL, MW and CSS; visualization, HXL and YW; supervision, ZWY and CSS; project administration, HXL and YW; funding acquisition, HXL and CSS. All authors have read and agreed to the published version of the manuscript. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Funding: This research was funded by the Medical Science and Technology Research Program Joint Construction Project of Henan Province (LHGJ20210003), Medical Science and Technology Research Program Joint Construction Project of Henan Province (LHGJ20220043), Science and Technology Development Planning Project of Henan Province (202102310499).

Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the by the Ethics Committee of the Henan Provincial People’s Hospital.

Informed Consent Statement: A waiver of informed consent was approved by the Ethics Committee of the Henan Provincial People’s Hospital since the waiver would not adversely affect the rights and welfare of the subjects because the study was retrospective and involved no more than minimal risk to the subjects.

Data Availability Statement: The authors confirm that the majority of the data supporting the findings of this study are available within the article. Raw data are available from the authors upon reasonable request. The data are not publicly available due to privacy restrictions.

Acknowledgments: The authors want to thank all the doctors, nurses, technicians and patients for their dedication to this study.

Conflicts of Interest: The authors declare no conflict of interest.

Disclosure: Part of the data was presented at the IDF World Diabetes Congress 2022, Lisbon and online, December 5-8.

International Diabetes Federation. IDF Diabetes Atlas, 10th edn. Brussels, Belgium: International Diabetes Federation, 2021.
National Collaborating Centre for Ws, Children's H. National Institute for Health and Care Excellence: Clinical Guidelines. Diabetes (Type 1 and Type 2) in Children and Young People: Diagnosis and Management. London: National Institute for Health and Care Excellence (UK).
Libman I, Haynes A, Lyons S, Pradeep P, Rwagasor E, Tung JY, et al. ISPAD Clinical Practice Consensus Guidelines 2022: Definition, epidemiology, and classification of diabetes in children and adolescents. Pediatric diabetes. 2022;23(8):1160-74.
ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 14. Children and Adolescents: Standards of Care in Diabetes-2023. Diabetes care. 2023;46(Suppl 1):S230-s53.
Tripathi A, Rizvi AA, Knight LM, Jerrell JM. Prevalence and impact of initial misclassification of pediatric type 1 diabetes mellitus. Southern medical journal. 2012;105(10):513-7.
Nieto T, Castillo B, Nieto J, Redondo MJ. Demographic and diagnostic markers in new onset pediatric type 1 and type 2 diabetes: differences and overlaps. Ann Pediatr Endocrinol Metab. 2022;27(2):121-5.
Dabelea D, Rewers A, Stafford JM, Standiford DA, Lawrence JM, Saydah S, et al. Trends in the prevalence of ketoacidosis at diabetes diagnosis: the SEARCH for diabetes in youth study. Pediatrics. 2014;133(4):e938-45.
Lévy-Marchal C, Patterson CC, Green A. Geographical variation of presentation at diagnosis of type I diabetes in children: the EURODIAB study. European and Dibetes. Diabetologia. 2001;44 Suppl 3:B75-80.
Cengiz E, Connor CG, Ruedy KJ, Beck RW, Kollman C, Klingensmith GJ, et al. Pediatric diabetes consortium T1D New Onset (NeOn) study: clinical outcomes during the first year following diagnosis. Pediatric diabetes. 2014;15(4):287-93.
Gungor N, Hannon T, Libman I, Bacha F, Arslanian S. Type 2 diabetes mellitus in youth: the complete picture to date. Pediatric Clinics. 2005;52(6):1579-609.
Umpierrez GE, Smiley D, Kitabchi AE. Narrative review: ketosis-prone type 2 diabetes mellitus. Ann Intern Med. 2006;144(5):350-7.
Hwang WB, Kim JH, Cho SM. Two cases of ketosis-prone diabetes mellitus in Korean adolescents. Ann Pediatr Endocrinol Metab. 2019;24(4):257-61.
Lu H, Hu F, Zeng Y, Zou L, Luo S, Sun Y, et al. Ketosis onset type 2 diabetes had better islet β-cell function and more serious insulin resistance. Journal of diabetes research. 2014;2014:510643.
Low JC, Felner EI, Muir AB, Brown M, Dorcelet M, Peng L, et al. Do obese children with diabetic ketoacidosis have type 1 or type 2 diabetes? Primary care diabetes. 2012;6(1):61-5.
Lebovitz HE, Banerji MA. Ketosis-prone diabetes (Flatbush diabetes): an emerging worldwide clinically important entity. Current diabetes reports. 2018;18:1-8.
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. Pediatric diabetes. 2018;19:155-77.
Vaibhav A, Mathai M, Gorman S. Atypical diabetes in children: ketosis-prone type 2 diabetes. BMJ case reports. 2013;2013.
Liu HX, Li GM, Zhou YW, Luo SH, Zheng XY, Yang DZ, et al. [Clinical characteristics and classification diagnosis of newly diagnosed diabetes onset with ketosis or ketoacidosis in adult patients]. Zhonghua yi xue za zhi. 2019;99(18):1369-74.
Thunander M, Törn C, Petersson C, Ossiansson B, Fornander J, Landin-Olsson M. Levels of C-peptide, body mass index and age, and their usefulness in classification of diabetes in relation to autoimmunity, in adults with newly diagnosed diabetes in Kronoberg, Sweden. European journal of endocrinology. 2012;166(6):1021-9.
Redondo MJ, Rodriguez LM, Escalante M, Smith EO, Balasubramanyam A, Haymond MW. Types of pediatric diabetes mellitus defined by anti-islet autoimmunity and random C-peptide at diagnosis. Pediatric diabetes. 2013;14(5):333-40.
Novac CN, Boboc AA, Nastac C, Balgradean M, Radulian G. Ketoacidosis Onset of Diabetes on a Patient with Normal C-Peptide Value. Maedica (Bucur). 2021;16(2):320-4.
Wang J, Zhang M, Liu Z, Wang X, Pang Y, Lu Y, et al. Heterogeneous clinical features of ketosis-prone type 2 diabetes mellitus patients: gender, age, loss of weight and HbA1c. Minerva endocrinologica. 2019;44(4):351-6.
Niechciał E, Rogowicz-Frontczak A, Piłaciński S, Fichna M, Skowrońska B, Fichna P, et al. Autoantibodies against zinc transporter 8 are related to age and metabolic state in patients with newly diagnosed autoimmune diabetes. Acta diabetologica. 2018;55(3):287-94.

No competing interests reported.

Download PDF

Reviews received at journal
13 Sep, 2024
Reviewers agreed at journal
07 Sep, 2024
Reviewers agreed at journal
05 Sep, 2024
Reviews received at journal
01 Aug, 2024
Reviewers agreed at journal
16 Jul, 2024
Reviewers agreed at journal
10 Jul, 2024
Reviewers invited by journal
10 Jul, 2024
Submission checks completed at journal
10 Jul, 2024
First submitted to journal
30 Jun, 2024

You are reading this latest preprint version

Practical Classification approach for newly diagnosed diabetes onset with ketosis or ketoacidosis in children and adolescent patients

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Highlights

Introduction

Materials and Methods

Participants

Data collection

Follow-up and classification

Statistical analysis

Results

Clinical presentations and distributions of included patients

Clinical characteristics between groups after 2 years

ROC curves and logistic regression analysis for classifying diabetes

Discussion

Conclusions

Declarations

References

Additional Declarations

Status:

Version 1