Factors related to glycemic control in children and adolescents with type 1 diabetes mellitus in Isfahan, Iran

Abstract

Purpose

Glycemic control is traditionally evaluated by monitoring of hemoglobin A1c (HbA1c). This study aims to explore factors related to glycemic control among pediatrics with type 1 diabetes mellitus (T1DM).

Methods

This cross‐sectional study was conducted among 454 students aged 6–18 years with T1DM in 2018. Demographic and disease related information were collected by a validated questionnaire. Generalized Linear Models (GLM) were used to investigate the association of explanatory variables with HbA1c concentration.

Results

The mean (SD) age of the participants was 11.7(± 3.3) years. The overall prevalence of suboptimal glycemic control was 85.5% (HbA1c ≥ 7%). Results showed that post pubertal children experienced a significant decrease in HbA1c levels compared to prepubertal children (β =—0.83, p = 0.003). Underweight children had an increase of 1.32% in HBA1c concentration compared with normal weight children (β = 1.32, P = 0.007). We also found that participants with passive smoking had higher HBA1c levels than those without (β = 0.536, P = 0.022).

Conclusions

The results indicated that age, BMI and passive smoking and were significantly associated with HbA1c levels. It is suggested that glycemic control is related to multiple factors and the interaction of these factors with each other may have positive or negative effects on it which should be investigated in future studies. Improved understanding in this area could lead to prevention of deterioration in glycemic control.

Introduction

Type 1 diabetes mellitus (T1DM) is a common chronic and metabolic disease in pediatric age [1].

The disease occurred by a deficit of insulin secretion and requires long-term follow-up and treatment for survival [2].

Based on the American Diabetes Association (ADA) guidelines for proper T1DM management and prevention of its related complications, the target level of HbA1C as a parameter of glycemic control is less than 7% [3–5]

However, achieving glycemic target is a major challenge in pediatric diabetes [6, 7]. Optimizing glycemic control is crucial for reducing life-threatening complications such as ketoacidosis and long-term complications such as retinopathy and neuropathy in children with T1DM [2, 8].

Evidence suggests that suboptimal glycemic control in the first months of diagnosis in childhood onset continues into later life [9–11].

Many studies have investigated factors related to proper glycemic control among children with T1DM in worldwide. Some of them reported an association between glycemic control and the type and dosage of insulin regimen [12, 13], family history of T1DM [14, 15], age [16–19], sex [20, 21], disease duration [13], frequency of self-monitoring of blood glucose and adherence to treatment [22]. Others did not demonstrate significant association between level of HbA1c and sex, type of insulin therapy, comorbidity with thyroid and celiac disease [16, 18].

Though some studies have examined factors that predict glycemic control in pediatric populations, but factors influencing glycemic control are poorly understood [12, 23]. There are limited studies in this field in Iran. However, evidences indicated suboptimal glycemic control in the majority of Iranian pediatric patients with T1DM [20].

Given that a better understanding of these factors could lead to better glycemic control, less disease related complications and designing more comprehensive educational and management protocols, the objective of this study was to identify the predictors affecting glycemic control in children and adolescents with T1DM in Isfahan.

Material and methods

Participants and study design

In this cross-sectional study, as a sub study of a National Institute for Medical Research Development (NIMAD), children and adolescents, aged 5–18 years, diagnosed with T1DM participated from schools of the urban and rural areas of Isfahan province in 2018. Inclusion criteria were: patients with T1DM and residence in Isfahan province, exclusion criteria were incomplete data on HbA1c level and refusal to participate in this research. A number of 454 students participated in the study by a convenient sampling scheme.

The study was approved by the General Directorate of Education of Isfahan province and NIMAD (research project number;958,386). Oral consent was obtained from the parents of students. This study was conducted with collaboration of Isfahan Endocrine and Metabolism Research Center, Isfahan, Iran.

Measurements

Demographic characteristics and T1DM related information were collected by a validated questionnaire according to parent report. The content validity of this questionnaire was examined by experts. After identifying the eligible students, parents of the students were invited to complete the questionnaire. Health teachers in schools of Isfahan province distributed the questionnaire among students. One parent from each student completed the questionnaire.

The questionnaire contained following parts; demographic characteristics (age, sex and place of residency), past medical history and clinical information including age at diagnosis, type of insulin regiment, comorbidity conditions (celiac disease, thyroid disease and hypertension).

Biochemical parameters included glycemic control) HbA1c) and triglycerides (TG) were also reported by parents.

Insulin regimens were classified into 6 categories: 1: Novorapid insulin plus NPH insulin, 2: regular insulin plus NPH insulin, 3: Levemir insulin plus Novorapid insulin, 4: Lantus insulin plus Novorapid insulin, 5: Novorapid insulin and 6: NovoMix insulin.

For passive smoking evaluation, the parents were asked to report whether your child lived with at least a smoker at home. The response to the question was “yes” or “no,”

Assessment of dietary patterns

Assessment of dietary habits was also achieved through parent report. Some dietary groups, including grilled fish, grilled chicken, grilled meat, fried hamburger, fried meat, fried chicken and fried fish, grilled corn and grilled potatoes were examined. The parents were asked to report how many times your child eat each of these foods (on usual consumption); the responses were ranked based on a five-point Likert scale with 1 = “never”, 2 = “rarely”, 3 = monthly, 4 = “weekly” and 5 = “daily”. Principal component analysis (PCA) was used to identify of dietary patterns. As a result, three dietary patterns including fried diet patterns, grilled meat consumption and grilled corn and potatoes were identified through three PCA. Then, the dietary patterns were categorized into two groups: lower consumption (values below the median) and higher consumption (values above the median).

Anthropometric evaluation

Body mass index (BMI) was computed by dividing weight (in kg) by height (in m²). Age- and sex-specific BMI percentile for all children were calculated according to the WHO growth curves.

BMI categorized as; BMI < 5^th percentile as underweight, BMI 5th to < 85th percentile as normal weight and BMI 85th to < 95th percentile as overweight and BMI ≥ 95th percentile as obese [24].

Optimal glycemic control was considered as HbA1c < 7% in individuals under 18 years of age as recommended by the ADA [3, 4]

Also, triglycerides (TG) levels above 110 mg/dL were considered as hypertriglyceridemia [25].

Data of Tanner stages were not available in the participants. Therefore, we defined the groups by age. The participants categorized into 3 groups: before puberty, during puberty and after puberty. The period of puberty was defined as 10.0–14.9 years of age for girls and 12.0–16.9 years for boys [26].

Statistical analysis

Quantitative variables were reported as mean and standard deviations (SD) and qualitative variables were reported as frequency and percentage. To compare quantitative and qualitative variables with diabetes control, independent t-test and chi square or fisher exact test (if applicable) were used, respectively. Then, the generalized linear models (GLM) were used to investigate the association of explanatory variables with risk of suboptimal glycemic control. HBA1C (Quantitative variable) was considered as the dependent variable in the models. The data were analyzed by the SPSS software version 18 (PASW Statistics for Windows, Chicago: SPSS Inc.). P value less than 0.05 was considered as significant.

Results

A total of 454 T1DM students aged 6 -18 years with mean age of 11.7 (3.3) were studied. 61.4% were female, 30.3% were overweight/obese and 4% were underweight.

The insulin regimens of most of the patients (86.9%; n = 285) was Lantus insulin plus Novorapid. Passive smoking was reported in 25.1% of the patients.

The prevalence of thyroid disease, celiac disease, hypertension, hypertriglyceridemia among the participants was 12.6%, 4.2%, 2.9% and 18.2%, respectively.

From initially included schoolchildren with T1DM in 345 cases the recorded HbA1c was accurate and reliable. So, their data were used for classification of patients with optimal and suboptimal glycemic control and 109 cases were excluded. Mean (SD) of HbA1c level in studied population was 8.62(1.81). 85.5% of patients had suboptimal glycemic control (HbA1c ≥ 7%). Mean (SD) of HbA1c in pre pubertal, pubertal and post pubertal groups were 8.63 (1.84), 8.98 (1.91) and 7.92 (1.31), respectively. Figure 1 shows that children 15 years or older have a significantly lower percentage of suboptimal glycemic control than children under 15 years.

Fig. 1.

HbA1c levels of children and adolescents with type 1 diabetes by age group

The association of the demographic and the disease-related characters with glycemic control are presented in Table 1. We have missing data for some variables due to the fact that the questionnaire was completed by parents. In spite of detailed explanations form parents, we have missing data as follows; residency status (32), BMI categories (111), age at diagnosis (48), hypertriglyceridemia (311) and HBA1C (109).

Table 1.

Characteristics of study population by HbA1c categories. Data are mean (± SD) or counts (%)

Characteristic		Total	Optimal glycemic control N (%) 50 (14.5%)	Suboptimal glycemic control N (%) 295(85.5%)	P value
Age, (year)*		11.7(3.3)	12.34(3.8)	11.3(3.2)	0.066
Pubertal classification	Pre puberty	107	14(13.1)	93(86.9)	0.024
	During puberty©	158	17(10.8)	141(89.2)
	Post puberty	80	19(23.8)	61(76.3)
Sex	Male	133	22(16.5)	111(83.5)	0.39
	Female	212	28(13.2)	184(86.8)
Residency	Isfahan city	136	21(15.4)	115(84.6)	0.626
	Other area	185	25(13.5)	160(86.5)
BMI, (kg/m2) *		19.9(4.4)	22.5(6.3)	19.4(3.8)	0.002
BMI categories	Underweight	12	1(8.3)	11(91.7)	0.194
	Normal weight	195	24(12.3)	171(87.7)
	Overweight	52	7(13.5)	45(86.5)
	Obese	38	13(34.2)	25(65.8)
Age of diagnosis, (year)*		7.1(3.7)	8.7(4.5)	6.8(3.4)	0.010
Age at diagnosis	< 12 years	284	32(11.3)	252(88.7)	0.002
	≥ 12 years	34	11(32.4)	23(67.6)
Family history of T1DM		40	5(12.5)	35(87.5)	0.708
Thyroid disease		39	6(15.4)	33(84.6)	0.974
Coeliac disease		13	0(0)	13(100)	0.228
Type of insulin used	Novorapid + NPH	10	0(0)	10(100)	0.192
	Regular + NPH	12	0(0)	12(100)
	Levemir + Novorapid	14	3(21.4)	11(78.6)
	Lantus + Novorapid	285	34(11.9)	251(88.1)
	Novorapid	4	1(25)	3(75)
	NovoMix	3	1(33.3)	2(66.7)
Fried food consumption	Low	145	30(20.7)	115(79.3)	0.027
	High	136	15(11)	121(89)
Grilled potatoes and corn consumption	Low	127	19(15)	108(85)	0.920
	High	169	26(15.4)	143(84.6)
Grilled meat consumption	Low	127	28(22)	99(78)	0.006
	High	150	15(10)	135(90)
Passive smoking	Exposed	84	9(10.7)	75(89.3)	0.241
Hypertriglyceridemia	< 110 mg/dL	117	19(16.2)	98(83.8)	0.766
	≥ 110 mg/dL	26	3(11.5)	23(88.5)
Diabetes duration (years)	< 5	153	27(17.6)	126(82.4)	0.101
	5–10	124	11(8.9)	113(91.1)
	> 10	41	5(12.2)	36(87.8)

BMI: Body mass index; Optimal glycaemic control; HbA1c < 7%, Suboptimal glycaemic control; HbA1c ≥ 7%

^*Mean (SD); P value based on t-test, other p values are based on chi-square test or Fisher exact test

© During puberty;10.0–14.9 years for girls and 12.0–16.9 years for boys

Mean (SD) age of T1DM patients with suboptimal glycemic control was younger than those with optimal glycemic control with a near significant P-value (0.066). BMI was significantly lower in participants with suboptimal glycemic control than in those with optimal glycemic control (P = 0.002).

Our results showed that the post pubertal participants (vs. prepubertal children) had a 0.83% decrease in HBA1c concentration (β = -0.83, P = 0.003).

We also found that underweight children had an increase of 1.32% in HBA1c concentration compared with normal weight children (β = 1.32, P = 0.007).

Moreover, participants with history of passive smoking had higher HBA1c levels than those without (β = 0.536, P = 0.022).

The children with coeliac disease had poorer glycemic control than those without comorbidity with a near significant P-value (0.062).

There were no significant differences in HbA1c among those with or without thyroid disease (β = 0.409, P = 0.185). Family history of diabetes was not associated with HbA1c levels (β = -0.084, P = 0.785) (Table 2).

Table 2.

Predictors of glycemic control in individuals with T1DM

		Crude models		Adjusted model*
Predictor		Beta	P-value	Beta	P-value
Age group (ref before puberty)	During puberty©	0.35	0.115	-0.138	0.573
	After puberty	-0.71	0.006	-0.829	0.003
Sex (ref male)	Female	-0.12	0.553	-0.022	0.918
BMI categories (ref normal weight)	Underweight	0.92	0.055	1.32	0.007
	Overweight	0.047	0.849	0.063	0.818
	Obese	-0.503	0.077	-0.237	0.423
Secondhand smoke exposure (ref unexposed)	Exposed	0.44	0.049	0.536	0.022
Grilled potatoes and corn consumption (ref low)	High	0.08	0.696	0.167	0.408
Grilled meat consumption (ref low)	High	0.24	0.255	0.293	0.184
Fried food consumption (ref low)	High	0.09	0.671	-0.024	0.916
Family history of T1DM (ref no)	Yes	0.12	0.695	-0.084	0.785
Thyroid disease (ref no)	Yes	0.42	0.167	0.409	0.185
Coeliac disease (ref no)	Yes	0.98	0.040	0.981	0.062

Ref: reference

^*Adjusted for all variables presented in this table

© During puberty;10.0–14.9 years for females and 12.0–16.9 years for males

Discussion

Our findings highlight associations between life stage, BMI and passive smoking and HbA1c levels. We found that post pubertal participants had significantly better glycemic control than prepubertal children.

Several studies showed a worsening of metabolic control during puberty [27–29]. In this context, some studies found an increase in HbA1c levels among adolescents aged 10–17 years with T1DM [30, 31]. It is may be due to hormonal and metabolic changes during puberty [32]. Evidence showed that insulin resistance rises during puberty, but reduce to prepubertal levels at the end of puberty [33, 34]. The rise of insulin resistance may be a risk factor for Type 1 diabetes [35, 36].

More appropriate glycemic control during post pubertal period may be due to that post pubertal participants may be more ready for successful independent self-management than prepubertal children and younger age group needs more supervision for proper control of blood glucose.

It may be also due to not enough education for both children and their parents at the time of diagnosis or lack of proper age related educations for children before and during puberty. Furthermore, fear from insulin injections may affect HbA1c levels [37]. Some studies have shown that fear from insulin injections was higher in younger children than older children [37, 38]. It is recommended to design more proper educational programs for children with T1DM and their parents for achieving better glycemic control as well as reducing periods of hypoglycemia.

In our participants’ mean HbA1c level) 8.62% (was similar to that found in the cross-sectional study on 1190 children and adolescents in New South Wales [39]. A study among 853 children with T1DM aged 0–18 years in Spain in 2017 showed that the mean of HbA1c was 7.3% [40].

Also, in the present study, 295 (85.5%) students had suboptimal glycemic control. Achievement of optimal glycemic control among children and adolescent with T1DM is a common problem. For example, a large international study on 44,058 T1DM patients aged < 15 years in developed countries showed that 15.7% to 46.4% of children had optimal glycemic control (HbA1c < 7.5) [41]. A study from Egypt has shown that 45.8% of children and adolescent with T1DM had suboptimal glycemic control [16]. In 2019, a study on 1095 children with T1DM aged 10–17 years showed that 35.8% of children had suboptimal glycemic control (HbA1c ≥ 9.5%) [42]. A study on 853 children less than 18 years old with T1DM in Spain showed that 66.6% of patients had optimal glycemic control [40]. A possible explanation for this variety in reported rate for glycemic control may come from using different cut-off point for suboptimal glycemic control definition [16, 42, 43]. So, the results cannot be compared with the findings of other studies.

The present study also showed that underweight children had higher HBA1c concentration compared with normal weight children. An international cross-sectional study, in 2018, among children aged 2–18 years showed that underweight and obese persons had poorer glycemic control than normal weight people [44]. A cross-sectional study (2015) in the United States, Germany and Austria showed that obesity in youth with T1DM associated with suboptimal glycemic control [45]. However, some studies didn’t find an association between BMI and HbA1c levels [16, 43, 46]. A cohort study on 635 children aged 7–24 years with T1DM was performed. There was no significant difference in HbA1c levels between overweight/obese and normal weight people [43].

Similar to other studies [16, 47, 48], we found that family history of diabetes had no effect on glycemic control. In 2015, a study among Kenyan children by Ngwiri et al. showed that family history of diabetes was not significantly associated with glycemic control [48]. A cross‑sectional study by Niba et al. among T1DM children in Cameroon shows that family history of diabetes was not significantly associated with glycemic control [47]. However, a previous Korean study showed that family history of diabetes was associated with high HbA1c in individuals with T1DM [49]. Riaz et al. stated that family history of diabetes was related to non-compliance to treatment and non-adherence to physical activity [50].

Evidence showed that 40% of children exposed to cigarette smoke, worldwide [51]. We found that 25.1% of participants had history of passive smoking at home. Since now a few studies have evaluated the relation between passive smoking and HBA1c in children with T1DM.

A study in Germany and Austria, on 27 561 patients less than 20 years old with T1DM showed that smokers display significantly poorer glycemic control than non-smokers [52]. We found that people with passive smoking had higher HBA1c levels than those without. This may be due to differences in lifestyle characteristics such as nutritional habits or physical activity. Previous studies suggested that the effects of passive smoking on HbA1c among adolescents and adults can depend on the nutrients (especially the consumption of omega-3 unsaturated fatty acids or vitamin C) [53, 54].

The differences mentioned between our study results and other studies from different populations may be due to differences in ethnic, cultural, geographical, health systems strategies in different countries as well as methodological factors.

To our knowledge, this study is the first study to evaluate factors related to glycemic control in children and adolescents with an age range of 6–18 years old from schools of the urban and rural areas of Isfahan province.

The major limitation of the present study was its cross-sectional design, which can limit a causal relationship. Data such as socioeconomic status, physical activity, sedentary behavior, sleep quality, other dietary components, frequency of blood sugar monitoring, adherence, history of previous hypoglycemic and dose of insulin were not available in the patient. In addition, laboratory data were not available for the diagnosis of DKA.

Information biases may have affected the results, because the data were collected using a parent-reported questionnaire. So, the results should be interpreted with caution.

Prospective studies are necessary to investigate the impact of modifiable life style-related variables, treatments and insulin dosage with the glycemic control in pediatric population with T1DM.

Conclusion

In summary, the results indicated that age, BMI and passive smoking were significantly associated with HbA1c levels. It is suggested that glycemic control is related to multiple factors and the interaction of these factors with each other may have positive or negative effects on it which should be investigated in future studies. Improved understanding in this area could lead to prevention of suboptimal glycemic control. Educational programs for children and their families are essential for optimal glycemic control in children, especially for younger children.

Funding

This work was funded by the National Institute for Medical Research Development (grant NO: 958386).

Declarations

Informed consent

Informed consent was obtained from all the patients.

This study was approved in ethics committee in National Institute for Medical Research Development (research project number;958386).

Conflict of interest

The authors declare that they have no conflict of interest.