Xue Zhao, Wei Gu

Endocrine Department of Children's Hospital Affiliated to Nanjing Medical University, Nanjing, Jiangsu, 210008

Keywords:Neonatal diabetes mellitus (NDM)Sulfonylureas Clinical characterization Growth and development

ABSTRACT Objective: We summarize the clinical and follow-up data of twenty-one children with neonatal diabetes mellitus (NDM) to strengthen the understanding of NDM and provide reference for clinical diagnosis and follow-up. Methods The clinical characteristics, growth and development of twenty-one children with NDM who were diagnosed and treated in the Children’s Hospital of Nanjing Medical University from January 2011 to August 2018 were retrospectively analyzed. Results The median age of diagnosis was 97 days and the follow-up period was 0.96 to 47.6 months years. At the time of new diagnosis, 7 cases were complicated with diabetic ketoacidosis and 3 cases with diabetic ketoacidosis. Seven patients had diabetic ketoacidosis (DKA) and three patients had diabetic ketosis (DK). Three cases were unclassified because of short follow-up time. Two patients are Transient neonatal diabetes mellitus (TNDM). Sixteen cases are Permanent neonatal diabetes mellitus (PNDM). Thirteen patients underwent drug-experiential treatment with a success rate of 53.8%. Twelve patients had growth and development disorders or language and motor retardation. Eleven cases were improved by genetic testing and the positive rate of gene mutation was 81.8%. There was no significant difference in treatment regimen, complications, genotype and other factors among different growth and development conditions (P ＞ 0.05). Fisher exact probability analysis of growth and development in different treatment schemes showed that there was no significant difference (P＞0.05). Conclusions Patients with KCNJ11 and ABCC8 gene mutations often have developmental disorders and sulfonylurea drugs are effective, which can improve the outcome of developmental disorders. There was no correlation between age, complications, genotype and the outcome of growth. When conditions permit, we should perfect gene detection as soon as possible to identify the type of mutation, guiding treatment and judging prognosis.

1. Introduction

Neonatal diabetes mellitus (NDM) is a special type of neonatal diabetes, which usually occurs within 6 months after birth. The main cause is single gene mutation. NDM is a rare disease. The latest data show that the incidence of this disease is 1, 000 and 160, 000.[1] The statistics of Japan and India in 2017 are one in 89000.[2，3] At present, there are no clear data and statistical results in China, but the incidence rate reported by each center is increasing year by year. According to the clinical outcome, it can be divided into to Transient neonatal diabetes mellitus (TNDM) and Permanent neonatal diabetes mellitus (PNDM). At present, more than 20 kinds of mutations have been identified, among which chromosome 6q24 mutation is a common cause of TNDM.[4] KCNJ11 or ABCC8 gene mutation (ATP sensitive potassium channel, K-ATP) is more common in PNDM.[5] Some K-ATP channel mutations can cause neurodevelopmental disorders, such as growth retardation, seizures or DEND syndrome. [6] In recent years, there is a new choice for the treatment of NDM-sulfonylureas, which is mainly manifested in the improvement of neurodevelopmental disorders in patients with NDM. In this study, the growth and development and drug use of 21 patients with NDM were retrospectively analyzed to explore the relationship between treatment and growth.

2. Materials and methods

2.1 Diagnostic criteria and exclusion criteria

Refer to the 2018 American ADA Diabetes guidelines [7]：Hyperglycemia persisted for more than 2 weeks within 6 months after birth, and NDM was diagnosed if infection, stress and druginduced hyperglycemia were excluded.

2.2 Object

Twenty-one patients with NDM diagnosed and treated by the Department of Endocrinology, Nanjing Children's Hospital from January 2011 to August 2018 were selected for follow-up. This study was approved by the Medical Ethics Committee of the Children's Hospital, and the informed consent form was signed by the patient's guardian.

2.3 Clinical data

The basic medical information of 21 patients with NDM, including sex, age of onset, height and weight, initial diagnosis, glycosylated hemoglobin, levels of insulin and C-peptide, islet cell autoantibody (ICA), insulin antibody (INS-Ab) and glutamic acid decarboxylase antibody (GAD-Ab), and the type of gene mutation, were recorded. The growth and development assessment includes height and weight (The reference standard is the height and weight standard table of nine major cities in China).

2.4 Statistical method

Stata12.0 was used to analyze the data statistically. The counting data are tested by Fisher exact probability method. The data of continuous variables with normal distribution and uniform variance between groups are represented by means ±SD and t-test is used. Non-normal distribution data are represented by Median (Q1-Q3). The Ranksum rank sum test is used to compare the continuous variables such as time, age and grade data which are not suitable for t-test. The two-sided test was used in all statistical tests, and the test level wasα = 0.05.

3. Result

3.1 General situation

3.1.1 The general condition of the patient

There were 4 girls and 17 boys in the 21 patients with NDM. The age of diagnosis ranged from 6 days to 202 days, the median time was 97 days, and the follow-up period was 0.96 to 147.6 months. The average HbAlc is 11.07%, with a maximum of 23.1% and 2 cases ＜7.0%. Seven cases had abnormal birth history, including intrauterine growth retardation, history of hypoxia asphyxia, premature infants and low birth weight. Seven cases were complicated with DKA and 3 cases with DK. One case was complicated with convulsion. There was no abnormality in lumbar puncture and cranial CT. EEG showed severe dysrhythmia, which was consistent with the characteristics of epileptic state. Brain MR indicates brain growth retardation and diagnosis of DEND syndrome. The highest blood glucose level of all patients after treatment was lower than that before treatment, which proved that the clinical treatment scheme was effective. The fasting serum C-peptide level of 1 patient was higher than the normal upper limit. With special faces and signs. Pituitary MRI showed slightly delayed myelination, widening of left temporal sublamellar space, T2 high signal in bilateral mastoid process and pituitary height about 2.5mm. Complicated with adrenocortical insufficiency, etc.

3.1.2 Diagnosis and treatment process

NDM patients with DKA or DK were treated with fluid replacement and continuous low dose insulin intravenously. The principles of treatment refer to the guidelines for diagnosis and treatment of diabetic ketoacidosis in children.[8] Transition from intravenous insulin infusion to subcutaneous insulin injection after remission of DKA. After communication, the parents of 13 patients signed an informed consent form and received glibenclamide treatment. The dose was gradually increased from 0.05～0.1mg/kg bid, and the insulin dose was reduced at the same time. At present, it is considered that when the maximum dose of 1mg/kg bid is reached or when the course of treatment is more than 4 weeks, [9] insulin has not been stopped, it will be regarded as a conversion failure. In this study, when the maximum dose was 1 mg/kg bid or the course of treatment was less than 3 weeks, the insulin was stopped and the blood glucose was normal, the conversion was successful. Finally, 2 cases were treated with glibenclamide combined with insulin, and 7 cases were treated independently with glibenclamide, of which 2 cases stopped taking glibenclamide more than 1 month and 8 months after treatment, and 1 case relapsed at the age of 9.7 years old.

3.1.3 Gene detection

Among the 11 cases of gene detection, there were 1 case of KCNJ11 and ABCC8 gene compound heterozygous mutation, 4 cases of KCNJ11 gene mutation, 1 case of FOXP3 gene mutation, 2 cases of INS gene mutation, 1 case of INSR gene mutation, and no NDM related gene mutation was found.

3.2 Follow-up

3.2.1 General situation

A total of 17 patients were followed up (≥1 time) for a followup period of 0.96 mi 147.6 months. There were 2 cases of TNDM, of which 1 case relapsed in childhood. Three cases were followed up for a short time without classification, and 16 cases had PNDM. Two cases were confirmed to die during follow-up, and all of themhad growth retardation and congenital abnormal development in the early stage. By the end of follow-up, there were 6 cases with complications：One case had neurological complications during the follow-up, combined with the previous history of meningitis, video-EEG and head MRI results, the diagnosis of epilepsy was clear. Antiepileptic therapy with 0.25g/d was given in 2014, followed by no seizures within 2 years. Follow-up of MRI revealed paraventricular leukomalacia, and no seizures occurred after followup. HbAlc fluctuated by 8.5%. Three cases were complicated with adrenocortical dysfunction and were still abnormal in late followup. One case considered the possibility of atypical congenital adrenocortical hyperplasia and was given low-dose hydrocortisone acetate during follow-up to recheck the normal range of ACTH. Combined with clinical manifestations and auxiliary examination, insulin resistance was confirmed after the gene test results of special facial cases were returned.

Table 1. Analysis of related factors in children with different growth and development（N=21）

3.2.2 Growth and development assessment

Most of the new cases are basically normal, or they are not evaluated because they are too young. During the late follow-up, 9 cases had normal growth and development, including 3 cases in the drug group. 12 cases had growth disorder or language and motor retardation, including 4 cases in the drug group. The height and weight of 3 patients in the insulin group were less than P10. The height or weight of 4 cases was less than P1 in the insulin group. One case was diagnosed as DEND syndrome, whose growth was obviously lagging behind, so he could not walk alone and could not call others.

3.2.3 Statistical analysis results

Sex, birth status, positive insulin antibody, choice of treatment, presence of DKA/DK, initial HbA1c and C-peptide of NDM type disease had no effect on growth and development. (P ＞0.05, Table 1) It was found that there was no correlation between different treatment schemes and growth and development by Fisher accurate probability analysis. (P＞0.05, Table 2)

Table 2. Accurate Fisher probability analysis of growth and development with different treatment schemes

4. Discussion

According to the clinical outcome, NDM can be divided into TNDM and PNDM. The onset age of TNDM patients is younger, the incidence of intrauterine growth retardation is higher. The initial dose of insulin in patients with DKA/DK is also relatively low. Most TNDM can be relieved within 1 year, but 60% of patients can relapse during puberty. If no remission was found in 18 months, it was PNDM.[10] In this study, there were only 2 cases of TNDM, 1 case of childhood recurrence and 1 case of normal child. Three cases with IUGR were PNDM. There was only one case of TNDM, and the reason for this difference may be related to the sample size. The K-ATP channel is an octamer structure consisting of four sulfonylurea receptor 1 (SUR1) subunits (ABCC8) and four Kir6.2 subunits (KCNJ11). When blood glucose increases, glucose enters β-cell metabolism to produce ATP. The ATP/ADP ratio changes and the K-ATP channel is closed. Changes in cell membrane potential activate calcium influx and release insulin.[11] When ABCC8 and KCNJ11 genes are mutated, the channel sensitivity changes, and some K-ATP channels are abnormally open, resulting in NDM.[12] Sulfonylureas mainly promote K-ATP channel closure by binding to SUR1 to release insulin. A 2014 Chicago study found : NDM subjects with 95% KCNJ11 or ABCC8 gene mutations could switch from insulin therapy to sulfonylurea therapy.[1]。A multicenter study conducted by Bowman Pamela et al[13] found that sulfonylureas can improve the neurodevelopmental outcome of NDM with channelrelated gene mutations, but this incomplete improvement will eventually reach a plateau. At the same time, this 10-year followup study also confirmed the safety and effectiveness of long-term treatment of sulfonylureas, and our study also found no adverse reactions other than digestive symptoms. A lot of studies have shown that the earlier the use of sulfonylureas, the more obvious the effect of sulfonylureas on nerve development, and the smaller the therapeutic dose required.[14，15] In addition, some NDM patients with non-channel related gene mutations such as chromosome 6q24 mutations are also responsive to sulfonylureas,[16，17] which may be related to the mechanism of drug binding to SUR1. K-ATP channels are also widely expressed in brain tissue. When the function is impaired, a series of neurodevelopmental disorders can occur, especially PNDM. The main clinical manifestations are attention deficit / hyperactivity disorder, sleep interruption, growth retardation and increased frequency of seizures.[11，18] David Carmody et al[6] conducted a comparative study on the neuropsychological and behavioral functions of NDM subjects with KCNJ11 mutations and their siblings. They found that mutant individuals without clinical symptoms may also have mild developmental disorders and their clinical manifestations range from mild delay without seizures to obvious developmental retardation, which adversely affects learning ability, social function and quality of life. The neurological and developmental disorders of NDM are not caused by diabetes, but are hereditary. In our study, the evaluation of neural development and growth and development is relatively simple, and the positive rate may be low. It is still necessary to find and develop more accurate and sensitive evaluation indicators in clinical work. In 1 case of NDM detected by gene, 5 cases (45.5%) had KCNJ11 or ABCC8 gene mutation. 4 cases were found to have growth and development disorders during follow-up. 80% of them were effective for sulfonylureas, including 2 cases of insulin combined with glibenclamide, which were basically consistent with the results of previous studies.[11] The remission time of repeated empirical treatment in 2 cases was much longer than 4 weeks, so it was inferred that prolonging the time of drug experimental treatment may improve the success rate. When conditions permit, gene detection should be improved as soon as possible to promote the conversion from insulin therapy to sulfonylurea therapy and improve the developmental outcome. A prospective study found that, the improvement of central nervous system function was the most obvious in PNDM patients with KCNJ11 gene mutation during the first year of sulfonylurea therapy.[19] In the early stage of treatment in our study, the improvement of growth and development disorder was more obvious, which may be related to the lack of nervous system development, and the catch-up increase of height and weight gradually slowed down with the increase of age. However, during the late follow-up, the lag of growth and development was aggravated in 2 cases. In our study, children with congenital developmental disorders or organ dysfunction are prone to multiple systemic developmental disorders. However, statistical results showed that genotypes, treatment regimens, NDM types, complications and other factors had no effect on growth and development in this study. NDM is a rare disease, the sample size is small, and the efficiency of statistical test is low, so the results of this study are not representative. Diabetic microvascular complications can occur in the early stage of PNDM. In this study, the three items of urinary microalbumin were negative, considering that the previous positive was related to infection. No other diabetes-related complications have been found. Some literatures have reported cases of NDM with hearing abnormalities or genetic metabolic diseases. In this study, patients with FOXP3 gene mutations also have hearing abnormalities. The correlation between clinical phenotypes and genotypes needs further study. In the study, it was also found that in one case of special face, the levels of insulin and C-peptide were significantly increased, and there was severe insulin resistance (insulin＞1000mU / L,C-peptide＞13.33nmol/L). The pathogenic gene is insulin receptor gene (INSR). The clinical phenotype of this gene can be divided into three types: Donohue syndrome (DS), Rabson-Mendenhall syndrome (RMS syndrome) and type A insulin resistance. Clinically, it can be accompanied by obvious growth and development lag, often accompanied by multiple systematic disorders. The diagnosis of this NDM requires vigilant screening for other special types of diabetes and genetic diagnosis is essential. At present, it is believed that the vast majority of antibody positive patients within 6 months are NDM, caused by FOXP3 gene mutation rather than type 1 diabetes.[18] However, in this study, there was only one patient with FOXP3 mutation and the antibody test was negative, considering the small sample size and detection error, and the positive antibody was not equal to type 1 diabetes. All the 7 children with abnormal birth history had growth retardation, suggesting that abnormal production history may be the risk factor for growth and development disorders in NDM. However, statistical analysis showed that P ＞0.05, there was no statistical difference. This may be related to the limitations of this study (the low efficiency of statistical test caused by the small sample size). During the follow-up, it was found that most of the HbAlc ＞7.5%. This is closely related to the low treatment compliance of parents and the inactive monitoring of blood glucose. The blood glucose management of NDM is still beset with difficulties.

In summary, conditions allow to improve gene detection as soon as possible and identify the type of mutation is of great significance for selecting treatment and judging prognosis. Sulfonylureas are effective in most NDM patients with ABCC8 and KCNJ11 gene mutations, and appropriately prolonging the time of experimental drug treatment may improve the success rate. In this study, there was no correlation between KCNJ11 or ABCC8 gene mutation, treatment regimen, complications and growth and development. Age, blood glucose, HbA1c and other factors had no effect on the choice of treatment. The clinical phenotype of NDM is closely related to genotypic heterogeneity, and there are significant individual differences, so it is necessary to establish a larger data management platform to provide guidance for clinical work.