To the Editor: Incidence of type 1 diabetes (T1D) in young children has been increasing over time in China. There are challenges in achieving metabolic control without leading to severe hypoglycemia or adverse effect on the quality of life in toddlers, as they showed marked sensitivity to insulin and higher variability in insulin requirements. Various insulin regimens are available for the treatment of T1D children through the flexible application of various insulin formulations. With changes in age, lifestyle, socioeconomic status, or eating habits, patients may switch the regimen as needed. The intensive insulin regimens, in which the insulin is required to be injected before each snack or meal, are not suitable for toddlers, who often have multiple meals. Continuous subcutaneous insulin infusion is limited in practical use due to poor economic condition and less imported machine. Therefore, twice-daily injection of self-mixing short and intermediate-acting insulin provides an alternative choice, which shows characteristics such as less number of injections, more accessible drugs, and multiple diets for toddlers. Previous studies showed that 75% of patients aged <10 years received conventional twice-daily therapy in New Zealand in 2017, while 39.92% of T1D children aged <5 years received twice-daily injection in our center in 2021.[3,4] This retrospective study performed comparison of different regimens in T1D toddlers during the first year after diagnosis.
This study respectively reviewed medical records of T1D patients receiving insulin injections between 2015 and 2019 in the department of endocrinology, genetics, and metabolism of Beijing Children's Hospital, Capital Medical University.
T1D was diagnosed based on clinical features, and diabetes-associated autoantibodies emerged in some patients. Clinical and genetic testing showed that the patients had no evidence of monogenetic diabetes or syndromic diabetes. All patients with diagnostic ages of 12 to 36 months received insulin injection immediately after diagnosis, adhering to the same insulin regimens for at least 12 months. We excluded patients with <12 months of follow-up and those who changed insulin regimen within 12 months. All parents had signed an informed consent form for using patients’ data. This study was approved by the ethics committee of Beijing Children's Hospital, Capital Medical University, Beijing, China (Approval No. 2020-k-180).
According to different insulin regimens, the patients were divided into two groups: group 1, in which the patients received twice-daily injections of self-mixing short and intermediate-acting insulin before breakfast and dinner; group 2, in which the patients received multiple daily injections (MDI) of rapid-acting insulin before each meal and long-acting analog once daily. The insulin regimen was voluntarily selected by parents based on the combination of family conditions, living habits, eating habits, patients’ personality, and doctors’ introduction. According to follow-up mode in our department (follow-up visit was made every 3 months, the patients additionally returned to the hospital at one month after discharge to assess adaption to situation), follow-up records at the 1st, 3rd, 6th, 9th, and 12th month of insulin treatment were summarized.
At each visit, diabetic patients underwent regular tests of glycated hemoglobin (HbA1c), fasting C-peptide (FCP) levels, fasting venous blood glucose, and urine routine test, and weight, height, insulin regimen, insulin dosage, self-reported frequency of hypoglycemia, frequency of self-glucose monitoring, ketosis, or symptomatic hypoglycemia were recorded. Patients’ data including diagnostic age, gender, weight, insulin dosage, frequency of hypoglycemia and frequency of self-glucose monitoring, HbA1c, FCP levels at baseline, and each follow-up visit (the 1st, 3rd, 6th, 9th, and 12th month) were extracted from medical records. Good glycemic control was defined as target HbA1c < 7.0%, and hypoglycemia was defined as a glucose value ≤ 3.9 mmol/L (70 mg/dL) according to International Society of Paediatric and Adolescent Diabetes (ISPAD) 2018 guidelines. Hypoglycemia was divided into two categories (glucose value: 3.7–3.9 mmol/L, ≤3.6 mmol/L). Severe hypoglycemic episodes were defined as coma, seizures, need for glucagon injections, or intravenous glucose infusion.
Descriptive statistics were calculated for demographic variables and changes in clinical variables from baseline to the 1st, 3rd, 6th, 9th, and 12th month using SPSS 24 (IBM Corp, Armonk, NY, USA). Analysis of variance was performed for continuous variables with normal distribution. Wilcoxon rank-sum test was performed for variables with non-normal distribution. P < 0.05 was considered statistically significant. To compare HbA1c and FCP levels at baseline between groups, mixed effect models were established using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). All data were analyzed to estimate the mean change at each time interval for each clinical outcome. Logistic regression analysis was performed to analyze the influences of group, gender, age, basal HbA1c, basal FCP, frequency of hypoglycemia, and frequency of self-glucose monitoring on the rate of achieving target HbA1c. Negative binomial regression models were used to assess the effects of different insulin regimens on hypoglycemia frequency after adjusting for gender, age, insulin dose, baseline HbA1c, FCP, and frequency of self-glucose monitoring.
From 2015 to 2019, 66 T1D patients with a diagnostic age of 12 to 36 months had been recorded, of whom, six was excluded, four was not followed up during the first year of treatment, one changed insulin regimen because neutral protamine Hagedorn (NPH) insulin was not available locally, and one died of an accident. Sixty patients were included in this study. There were 34 patients in group 1 and 26 patients in group 2. Baseline characteristics of subjects were shown in Supplementary Table 1, https://links.lww.com/CM9/B187. The ethnic origins of all patients were Chinese; their mean diagnostic age was 23.9 ± 6.1 months. Mean HbA1c before insulin treatment was 10.67 ± 1.77%, and mean FCP before insulin treatment was 0.32 ± 0.25 ng/mL (median: 0.27 ng/mL, range < 0.01–1.3 ng/mL). There were no statistically significant differences in age, sex, baseline HbA1c, FCP level, and insulin dose between the two groups [Supplementary Table 1, https://links.lww.com/CM9/B187].
HbA1c level showed a downward trend over time in both groups and was decreased faster in group 1 [Supplementary Table 1, https://links.lww.com/CM9/B187]. HbA1c level was statistically lower in group 1 at the 3rd month (P = 0.036), while no significant difference in HbA1c level was detected at other time points. According to parametric estimated values, the decrease in HbA1c level was gradually accelerated within the first 6 months in both groups; the interaction analysis showed that the interaction effects between two groups were statistically significant at the 3rd month (P = 0.0032), indicating a more obvious decreasing trend of HbA1c level in group 1 within the first 3 months, with a smaller difference in decreasing trend of HbA1c level at other time points.
During the first year, the total rate of achieving target HbA1c was 85.29% in group 1 and 69.23% in group 2, indicating no statitically significant difference (P = 0.134). The logistic regression analysis showed that both insulin regimen and average frequency of hypoglycemia exerted no significant effect on the rate of achieving target HbA1c, only the diagnostic age had a significant effect on the rate of achieving target HbA1c (odds ratio [OR] = 0.844; 95% confidence interval [CI] 0.726–0.983; P = 0.029).
Changes in FCP from baseline to the 1st, 3rd, 6th, 9th, and 12th month were shown in Supplementary Table 1, https://links.lww.com/CM9/B187. As many FCPs were lower than 0.01 ng/mL (below the detection limit), a rank-based mixed-effect model was established to analyze FCP. After adjusting for gender, age, and insulin dose, there was no significant difference in FCP between the two groups (P = 0.073). FCP was decreased gradually in both groups, and interaction effect analysis indicated no significant difference in the decreasing trend between two groups. There was no difference in the rate of decline in FCP between two groups.
The frequency of self-glucose monitoring was categorized as <4, 4–6, >6 times/day or continuous glucose monitoring (CGM). Fifty-eight percent (35/60) of patients received blood glucose monitoring (BGM), 35.0% (21/60) of patients received CGM, and 21.6% (13/60) of patients initially received BGM and then switched to CGM in the follow-up period [Supplementary Table 1, https://links.lww.com/CM9/B187]. The frequency at each time point was analyzed, indicating no significant difference between two groups.
No severe hypoglycemic episodes occurred in both groups. The patients in both groups had significant hypoglycemia, with a median frequency of hypoglycemia of 14 times/month in group 1 and 8.5 times/month in group 2 [Supplementary Table 1, https://links.lww.com/CM9/B187]. Hypoglycemia was divided into two categories (glucose value: 3.7–3.9 mmol/L and ≤3.6 mmol/L). At the 1st, 3rd, 6th, and 9th month, the frequency of hypoglycemia showed no statistically significant difference between two groups. At the 12th month, the frequency of hypoglycemia was higher in group 1, with a median of 16 times/month vs. 8 times/month in group 2 (P = 0.025). The incidences of two categories of hypoglycemia were compared at the 12th month between two groups, the incidence of glucose value ≤3.6 mmol/L was higher in group 1 than in group 2 (P = 0.042), whereas there were no differences in the incidence of glucose value of 3.7 to 3.9 mmol/L between two groups.
Considering that factors such as the frequency of self-glucose monitoring may affect the frequency of hypoglycemia, a multilevel negative binomial model was applied. As there were a lot of missing data on hypoglycemia at the 12th month, no sufficient data at the 12th month were input into the multilevel negative binomial model, only the data at the first four time points (the 1st, 3rd, 6th, and 9th month) were analyzed. Adjusting for gender, age, insulin dose, and frequency of self-glucose monitoring, there was no statistically significant difference in the frequency of hypoglycemia between the two groups (P = 0.536). The data of glucose value of 3.7 to 3.9 mmol/L and glucose value ≤3.6 mmol/L were analyzed separately, indicating no significant difference between two groups).
Because the frequency of hypoglycemia after discharge was self-reported, we also collected the blood glucose data detected during hospitalization, the frequency of self-glucose monitoring was 7 times/day in all patients, meanwhile the diet and exercise were controlled, the medical records on glucose value were clear. There were no significant differences in hospitalization days and hypoglycemia frequency between the two groups [Supplementary Table 1, https://links.lww.com/CM9/B187].
To evaluate the effectiveness of twice-daily injections and MDI of insulin in T1D toddlers, we analyzed the rate of achieving HbA1c <7.0% to evaluate glycemic control. The rate of achieving target HbA1c was essentially the same in both groups (85.29% and 69.23%), and the logistic regression analysis revealed no correlation between the rate of achieving target HbA1c and the insulin regimen, this is in line with the previous finding which revealed no correlation between HbA1b value and insulin regimens.[3,5] The rate of achieving target HbA1c in this study is relatively higher than the reported rates of 17% to 49% in high-income countries, but it is in line with the reported rate of 56.3% (228/405) in T1D children in our center in 2019. Additionally, in this study, HbA1c levels in group 1 decreased faster in the first 3 months. A single-center study including 69 Egyptian toddlers with T1D showed that NPH group achieved a lower HbA1c level compared with the long-acting analog group, which is partially in line with our findings that HbA1c level decreased faster in the group receiving twice-daily insulin injection. However, in both studies, there was no evidence that the rapid decline of HbA1c was directly related to the frequency of hypoglycemia, consisting with the finding that the frequency of hypoglycemia was not directly associated with the type of regimen or the number of injections. In this study, there were no statistically significant difference in the frequency of hypoglycemia at the 1st, 3rd, 6th, and 9th month between two groups, indicating that the potent glucose-lowering effect of twice-daily injection regimen is not a direct result of hypoglycemia. At the 12th month, the hypoglycemia frequency was higher in group of twice-daily regimen, but this could not be further proved by a multilevel negative binomial analysis due to missing data on hypoglycemia at this time point, it is difficult to determine the main influence factor. More detailed glucose records are needed to be analyzed in the future.
This retrospective study has certain limitations. Some missing values in medical records at every time point affected the effect evaluation. The frequency of blood glucose and hypoglycemia monitoring was self-reported, which was not very objective. In addition, a detailed blood glucose level record reflecting glycemic variability is needed. A randomized controlled trial with a longer follow-up period will be performed in the future.
In conclusion, within the first year of diagnosis, the metabolic control of twice-daily injection of self-mixing short and intermediate-acting insulin is similar to that of MDI, and the twice-daily insulin injection does not significantly increase the frequency of hypoglycemia in the first 9 months. According to ISPAD 2018 guidelines, whatever insulin regimen is chosen, it must be supported by comprehensive education appropriate for the age, maturity, and individual needs of the child and family (Grade A). This study proves that T1D toddlers after initial diagnosis can be treated with twice-daily insulin injection regimen if warranted by the specific case or family situation.
This work was supported by a grant from the Beijing Municipal Science & Technology Commission (No. Z201100005520061).
Conflicts of interest
1. Fu JF, Liang L, Gong CX, Xiong F, Luo FH, Liu GL, et al. Status and trends of diabetes in Chinese children: analysis of data from 14 medical centers. World J Pediatr 2013;9:127–134. doi: 10.1007/s12519-013-0414-4.
2. Dovc K, Boughton C, Tauschmann M, Thabit H, Bally L, Allen JM, et al. Young children have higher variability of insulin requirements: observations during hybrid closed-loop insulin delivery. Diabetes Care 2019;42:1344–1347. doi: 10.2337/dc18-2625.
3. Miao Q, Chunxiu G, Bingyan C, Rui W, Guoshuang F, Di W, et al. Frequency of SMBG correlates with HbA1c and acute complications in children and adolescents with type 1 diabetes. Chin J Diabetes Mellitus 2021;13:462–469. doi: 10.3760/cma.j.cn115791-20200921-00575.
4. Selvakumar D, Al-Sallami HS, de Bock M, Ambler GR, Benitez-Aguirre P, Wiltshire E, et al. Insulin regimens for newly diagnosed children with type 1 diabetes mellitus in Australia and New Zealand: a survey of current practice. J Paediatr Child Health 2017;53:1208–1214. doi: 10.1111/jpc.13631.
5. Gong CX, Wei LY, Wu D, Cao BY, Meng X, Wang LL. Effectiveness of multiple daily injections or continuous subcutaneous insulin infusion for children with type 1 diabetes mellitus in clinical practice. Int J Endocrinol 2014;2014:526591. doi: 10.1155/2014/526591.
6. Charalampopoulos D, Hermann JM, Svensson J, Skrivarhaug T, Maahs DM, Akesson K, et al. Exploring variation in glycemic control across and within eight high-income countries: a cross-sectional analysis of 64,666 children and adolescents with type 1 diabetes. Diabetes Care 2018;41:1180–1187. doi: 10.2337/dc17-2271.
7. Hassan MM, Arafa N, Abdou M, Hussein O. Characteristics of diabetes diagnosis and control in toddlers and preschoolers from families with limited resources: a single center experience. Diabetes Res Clin Pract 2020;159:107966. doi: 10.1016/j.diabres.2019.107966.
8. Shalitin S, Phillip M. Which factors predict glycemic control in children diagnosed with type 1 diabetes before 6.5 years of age? Acta Diabetol 2012;49:355–362. doi: 10.1007/s00592-011-0321-x.