The type 1 diabetes (T1D) is a chronic disease that poses a challenge to public health, worldwide. According to the 2024 International Diabetes Federation (IDF) reports, among 9.15 million people living with T1D globally, 1.81 million (19.8%) are younger than 20 years.1 The T1D leads to complications like cardiovascular dysfunction, dyslipidaemia, albuminuria, end-stage renal disease and poor quality of life.2–4 Moreover, life expectancy of an African child diagnosed with T1D at the age of 10 varies around 19 years, which means 9 years after diagnosis.1 To prevent complications, glycosylated haemoglobin (HbA1c) level must be below 6.5% according to the IDF.1,5
In fact, an adequate diet helps to maintain HbA1c below 6.5%, while children are followed with insulin.6 This diet must correspond to energy needs as defined by American Diabetes Association (ADA).7 Also, it must be balanced and diversified.8,9 According to the International Society for Paediatric and Adolescent Diabetes (ISPAD) 2022, the diet is balanced when calories come from 40-50% of carbohydrates, 15-25% of protein and 30-40% of fat.10
Previous studies in Africa have shown that the majority of children with T1D have mean HbA1c over the threshold.11,12 According to the caloric distribution, a study conducted by Pillay in 2009 among 30 children with TID in KwaZulu-Natal, South Africa, found that calories came from 52% carbohydrate, 16% protein, and 32% fat.12 Another study conducted in Newcastle, Australia, by Seckold in 2019 among 24 children revealed that calories came from 48% carbohydrate, 16% protein, 33% lipid and children’s HbA1c was not correlated with daily carbohydrate, protein or lipid intake.13 Furthermore, a single-centre study in children with T1D conducted by Gilbertson in 2018 in Victoria, Australia, found a distribution of 49% carbohydrate, 19% protein and 33% lipid.14
In Kinshasa, Democratic Republic of Congo (DRC), Almost all children with T1D have HbA1c levels over the threshold, despite all efforts on treatment.15 Actually, there is no explanation on that from the literature. In this study, we wanted to verify whether there is a link between high level of HbA1c and dietary intake among children. The results could give orientation in the treatment of children followed up in diabetic clinics in Kinshasa.
METHODS
This was a multicentre analytical cross-sectional study carried out from March to May 2024. The study was conducted in the six clinics for diabetic children in Kinshasa (Table 1). These clinics provide free care for children with diabetes. Inclusion criteria included healthcare centres recognised as clinics for diabetic children and those equipped with a device for measuring HbA1c in children. The exclusion criterion was any diabetic clinic that did not measure HbA1c in children during the study period.
Study Population and Sampling
The study was conducted among 113 children under 19 years old, living with T1D, out of 139 who came to medical appointments.
-
Inclusion criteria: children who were followed up for T1D three months before the study, who performed HbA1c and whom parents or guardians informed consent.
-
Exclusion criteria: children who did not remember foods they ate one day before the appointment, who had vomiting and diarrhoea 24 hours before, and who were not compliant in insulin treatment recommendations.
The selection was exhaustive. It took into account all children who met selection criteria. We had 38 from Boyambi, 11 from Binza, 11 from Lunda, 7 from Etonga, 14 from Mont-Amba, and 32 from 2nd Street Limete.
Data Collection and Variables
The data collection technique used was the interview with questions on sociodemographic characteristics and 24-hours recall. The first was done in person on the day of the appointment. The questions were asked to the parents or guardians who accompanied the children under 12 years old. For older children who came alone to the appointment, questions were asked to them, then to their parents by phone calls for confirmation. The second interview was done by a phone call on another non-consecutive day.
Glycosylated haemoglobin (HbA1c)
The HbA1c, dependent variable, was obtained by placing children blood in the “Hemocue HbA1c” device. Each clinic had the same type of device. The results given by the device did not exceed 14%. All results more than 14% were displayed as “>14%”. Results were noted in the card of each child by laboratory technicians. The investigators were picking the results up from cards. HbA1c threshold was 6.5%.1
Dietary intake
The two dietary intake variables, caloric consumption and dietary balance, were independent variables. Caloric consumption was considered normal when children had energy consumption corresponding to their needs.7 Balanced diet, when energy provided from 45-60% carbohydrate, 15-20% protein and 25-35% lipid.10 Dietary intake was known from the 24-hours recall. We used the average of two recalls from two non-consecutive days, one for Sunday and another for a weekday. The IF1976 precision gram scale, model 14191-2103B, was used to weigh foods reported on 24-hour recalls. These foods came from markets and restaurants around health centres. The food composition table used was that of the United State Department of Agriculture (USDA),16 supplemented by that of traditional foods of the DRC.17
Sociodemographic variables
We collected age and sex of children, type of family and main activity of parents or guardians.
Data Analysis
The analysis was performed on SPSS v26 software. Qualitative data were summarized as absolute frequencies and percentages. Quantitative data were summarized as mean and standard deviation (SD) for normally distributed variables and as median and interquartile range (IQR) when distribution was not normal. HbA1c was categorized in low (HbA1c < 6,5%) and high (HbA1c ≥ 6,5%). Caloric consumption was categorized in low, normal and over consumption. Dietary balance was categorized into balanced and unbalanced diet. The comparison of caloric consumption in the two groups of children according to HbA1c, was made by Mann-Whitney U test because the distribution was not normal. Then associations were tested by chi-square test at 5% significance level between categorical variables, dietary intake variables and HbA1c.
RESULTS
Among 113 children, 50% had a daily consumption between 1539 and 2624 Kilocalories (Kcal). Children had more carbohydrate (273g, IQR=184) in their daily meals than lipid and protein. Energy intake came from 55% of carbohydrate, 13% of protein and 32% of lipid. About 89.4% of children had HbA1c over 6.5%, 54.9% had a low daily caloric intake, and 80.5% consumed an unbalanced diet. Children with low HbA1c levels had consumed as the same calories as children with high HbA1c levels. Moreover, there was no significant association between caloric intake or dietary balance and HbA1c levels (p = 0.316 and p = 0.303, respectively).
Sociodemographic characteristics
The 113 children included had a median age of 15 (IQR=5) years. The majority of children (n=84, 74.3%) were over 13 years old. Boys were 61 (54%) and girls 52 (46%). About half (n=55, 48.7%) were living in a two-parent family, and almost all parents or guardians (n=9, 92%) had remunerated employment.
Dietary Intake
The Table 2 shows the amount of each macronutrient consumed in grams, as well as the energy consumed produced by these macronutrients. According to the Table 2, 50% of children had a daily consumption varying between 1539 and 2624 Kcal. About 25% had an energy consumption greater than 2624 Kcal, and 25% less than 1539 Kcal. For macronutrients, children had more carbohydrate (273 g, IQR=184) than other energy nutrients. The Table 3 shows that the energy intake came from 55% of carbohydrate, 13% of protein and 32% of lipid.
Energy Intake and HbA1c Values
Only 12 (10.6%) children had HbA1c less than 6.5% and 101 (89.4%) had HbA1c greater than or equal to 6.5%. The Table 4 shows the comparison of the energy that children consumed in the two groups of children according to their HbA1c values. According to the Table 4, the amount of calorie and that of macronutrients consumed by children with low HbA1c and by those with high HbA1c are statistically the same (p > 0.05). The Table 5 shows the associations of dietary intake variables and HbA1c. As illustrated in the Table 5, over half of the children (54.9%) had a low daily caloric intake, and 80.5% consumed an unbalanced diet. However, statistical analysis showed no significant association between caloric intake or dietary balance and HbA1c levels (p = 0.316 and p = 0.303, respectively).
DISCUSSION
Diabetes mellitus in children is poorly addressed. In this study, about 89.4% of children had HbA1c over 6.5%. This implies that almost all children were at risk of developing complications related to T1D.2–4 This is similar to some African studies. Although they used higher HbA1c thresholds, Niba in Cameroon in 2017 with 7.5% and Pillay in South Africa in 2009 with 8%, they have found respectively 76% and 83% of children with high HbA1c levels.11,12 All these studies had been conducted in Africa, which can be the reason of similarity due to the context of poverty.
Moreover, about 54.9% of children had a low caloric intake, and 80.5% consumed an unbalanced diet. This could be justified by the poverty in the Kinshasa population.18,19 Furthermore, energy intake came from 55% of carbohydrate, 13% of protein and 32% of lipid. By comparing with the ISPAD recommendations, which determine the balanced meal of 40 to 50% carbohydrate, 15 to 25% protein and 30 to 40% fat,10 meal of children in this study was more rich in carbohydrate than in protein. The mean protein intake was lowest compared to the studies of Pillay in 2009 in Kwazulu-Natal, South Africa12; Seckold R in 2019 in Newcastle, Australia13; and Gilbertson HR in 2018 in Victoria, Australia.14 This difference could be explained by the lowest socioeconomic conditions in Kinshasa, which has an influence on the quality of foods.18
According to the Mann Whitney U test, there was no difference in the amount of calories and macromolecules consumed by children with low HbA1c and those with high HbA1c. There was, also, no association between caloric intake or dietary balance and HbA1c. These results showed that what children used to eat did not influence their HbA1c levels. The result is consistent with a study conducted by Seckold R in 2019 in Newcastle, Australia, where HbA1c was not correlated with daily carbohydrate, protein or fat intake (p>0.05).13
Strengths
Unlike studies on the nutrition of diabetic children conducted in other countries, this study was multicentre. The use of exhaustive sampling in the six diabetic clinics in Kinshasa allowed to increase the sample size.
Limitations
The 24-hour recall presented the risk of information bias, such as memorization and instrument bias. Regarding memorization bias, children and parents or guardians could forget a detail of a dish. Also, recalling a single day could not justify their usual diet. To minimize this bias, we conducted 24h-hours recall on two non-consecutive days and reported the average. During each recall, we asked questions to children and verified them to their parents or guardians. Regarding instruments, we used food composition tables and a scale to weigh foods. Nutritional values of some foods vary from one table to another, and there is no table that includes everything that is consumed locally in Kinshasa. Thus, we used USDA as the main table and local foods table from Mbemba as a secondary table.17 It was also difficult to assess the quantity and to weigh certain foods revealed during the 24-hour recall. For this, we had carried out the reconciliation of weights with some similar foods whose weights were already known.
CONCLUSIONS
The persistent HbA1c high levels in children with T1D is a problem in Kinshasa. The findings highlight the need for comprehensive strategies to manage T1D in children, emphasizing that dietary intake alone may not be sufficient to control HbA1c levels. There is a clear need for further research to explore non-nutritional factors affecting glycaemic control, such as socioeconomic status, environmental aspects, access to healthcare, and education on diabetes management. The study underscores the importance of advocacy for improved healthcare resources and education for families managing T1D in low-resource settings.
Acknowledgements
We are grateful to academic authorities from Kinshasa School of Public Health and authorities from the six diabetic clinics where the study has been conducted for facilitating the data collection.
Ethics statement
This study was approved by the ethics committee of the Kinshasa School of Public Health, University of Kinshasa, under approval number ESP/CE/32/2024 dated February 14, 2024. Informed consent was obtained from all participants involved in this study.
Funding
This research received no external funding
Authorship contributions
Stany Gbiambi has designed and conducted this study from the beginning to the end. Marie-Claire Muyer monitored technical aspects and standards of scientific research by correcting several times until we had the final version to submit.
Disclosure of interest
The authors completed the ICMJE Disclosure of Interest Form (available upon request from the corresponding author) and disclose no relevant interests.
Correspondence to:
Gbiambi Legbia Stany
Department of Nutrition, Kinshasa School of Public Health
22/B Q. Lokoro, Matete Township, Kinshasa
Democratic Republic of the Congo
legbiastany@gmail.com