Since the publication of the Diabetes Control and Complications Trial, HbA1c has been considered the main parameter for assessment of metabolic control in both children and adults with type 1 diabetes (T1DM).1 However, recent studies with continuous glucose monitoring (CGM) systems have revealed wide fluctuations in glucose values in children with T1DM, even in those with excellent HbA1c levels.2,3 Interest in studying the impact of these glucose fluctuations has therefore increased.4–6 The concept of glycaemic variability encompasses both diurnal variability (fluctuations in glucose within a single day) and day-to-day variability (fluctuations in glucose between one day and the next). Diurnal variability can be estimated using the standard deviation (SD) around the mean of glucose values for one day, and through the mean amplitude of glycaemic excursions (MAGE).7 The mean of daily differences (MODD)8 is a day-to-day variability marker. Other parameters for evaluating the risk of reaching extreme glycaemic values are the low blood glucose index (LBGI), which is a measure of the risk of hypoglycaemia,9 and the high blood glucose index (HBGI), which is a measure of the risk of hyperglycaemia.10

Glycaemic variability has recently been associated with an increase in the production of free radicals and oxidative stress.11 Isoprostanes are considered good markers of oxidative stress, as they are stable products derived from the process of oxidation of arachidonic acid by oxygen free radicals and have been shown to increase with oxidative damage.12,13 Isoprostanes can be estimated by measuring the 24-h urinary excretion rate of 8-iso-prostaglandin F2 alpha (8-iso-PGF2α).

The objectives of this study are to assess glycaemic variability and oxidative stress by the urinary excretion rate of 8-iso-PGF2α in a group of children and adolescents with T1DM attending a summer camp, and to analyse the relationship between both.

Table 1 shows the sociodemographic and clinical characteristics of the participants. As regards treatment, 3 of the children had CSII pumps with a rapid-acting insulin analogue, and the remaining children used MDII. Of the latter, 45 received a combination of long-acting insulin analogues (insulin glargine or detemir) and rapid-acting insulin analogues (insulin lispro, aspart or glulisine), and 6 received a combination of neutral protamine Hagedorn (NPH) insulin and rapid-acting insulin analogues (insulin lispro, aspart or glulisine). None of the participants exhibited microvascular complications.

A significant reduction in insulin requirement was observed during the camp (0.85±0.26IU/kg at the start of the camp vs 0.69±0.24IU/kg at the end, p<0.0001). Glycaemic control parameters, markers of glycaemic variability and episodes of acute diabetes-related complications are shown in Table 2. The mean number of episodes of mild hypoglycaemia per subject during the whole camp was 5.1 (range 0–17). One child had an episode of severe hypoglycaemia requiring intramuscular administration of glucagon to resolve it. Five episodes of ketosis were identified; these were treated at the camp and none of the cases developed into ketoacidosis. Of all children, 54% exhibited an HbA1c value ≤7.5% (58mmol/mol). The median SD, MAGE and MODD indexes were in the high range, mean LBGI was within the moderate risk category and mean HBGI was within the low risk category. 74% of the children had a MODD value of over 36mg/dl. LBGI values indicated a high risk of hypoglycaemia in 26% of the participants and a moderate risk in 39%. The median urinary excretion rate of 8-iso-PGF2α measured in 14 children at the end of the camp was 864.39pg/mg creatinine.

We found no statistically significant differences in markers of glycaemic variability (SD, MAGE, MODD and HBGI), HbA1c, duration of diabetes (4.79±3.28 years vs 4.75±3.18 years; p=0.98), age (11.86±1.91 years vs 11.25±2.46 years; p=0.33) and sex (50% male vs 42.5% male, p=0.63) between the children who submitted the 24-hour urine sample and those who did not (Table 2). LBGI and time in hypoglycaemia were significantly lower in the group of children who did not submit the urine sample (Table 2).

Table 3 shows the positive and statistically significant correlations that were observed between duration of diabetes, HbA1c, mean glycaemia, SD and the HBGI, and MAGE and MODD indices. LBGI correlated positively with MAGE and negatively with mean glycaemia. However, neither age nor urinary excretion rate of 8-iso-PGF2α correlated with markers of glycaemic variability (Table 3).

This study shows considerable glycaemic variability in children and adolescents with T1DM attending a summer camp. However, no correlations were found between high glycaemic variability and oxidative stress measured through the urinary excretion rate of 8-iso-PGF2α.

The objectives of the study included, on the one hand, assessment of glycaemic variability in children with T1DM, specifically in the context of summer camp. It is known that children and adolescents with T1DM are very prone to exhibit extreme glycaemic values, mainly due to variations in insulin sensitivity, level of physical activity and food intake.16,17 Specialised camps for children and adolescents with T1DM offer us an opportunity to study glycaemic excursions outside the hospital setting.18 For example, Choleau et al.19 assessed day-to-day glycaemic variability, estimated using the MODD index, which was calculated from blood glucose readings, in a group of 100 children with T1DM treated with MDII at a summer camp, reporting that the median MODD index was 77mg/dl and that in 99% of the participants the MODD value was over 36mg/dl. Another recent study20 of 48 children and adolescents with T1DM who underwent an ambulatory 3-day CGM showed high day-to-day glycaemic variability (MODD 63mg/dl in the group of children on MDII and MODD 61mg/dl in the group on CSII), and a high diurnal variability, especially in the group of children treated with MDII (MAGE 117mg/dl in the group of children on MDII and MAGE 88mg/dl in the group on CSII). Our results for glycaemic variability are consistent with data presented previously, and are even higher than those found in studies assessing children in their everyday lives, highlighting the fact that summer camps, even under the supervision of a specialist medical team, constitute an unusual situation.

On the other hand, we wanted to study oxidative stress and its relationship with glycaemic variability. Cerriello and Ihnat demonstrated that glycaemic variability can contribute to accelerated formation of free radicals,21 and furthermore Giacco et al.22 showed that oxidative stress can play an essential role in the development of diabetes-related microvascular and macrovascular complications, so this evidence indicates that glycaemic variability could have a greater deleterious effect on the development of complications than sustained hyperglycaemia.23 Nevertheless, the debate on the significance of glycaemic variability as a clinical finding in diabetes continues. Indeed, a recent study carried out on children with T1DM found no relationship between glycaemic variability assessed with CGM and vascular dysfunction.24 Monnier et al.25 demonstrated a close relationship between glycaemic variability and oxidative stress measured through urinary excretion rate of 8-iso-PGF2α in 21 subjects with type 2 diabetes (T2DM). However, it is difficult to replicate these results. DeVries's group, was not able to demonstrate this association in a group of adult patients with T1DM,26 nor in another group of patients with T2DM treated with oral antidiabetics (OAD).27 One possible explanation for the inconsistency of these results is that the methodology used to measure the urinary excretion rate of 8-iso-PGF2α differed between the two groups, as DeVries26,27 used tandem mass spectrometry, which is less prone to interference than the enzyme immunoassay technique used by Monnier.14 Another factor that could exert an influence is the potential effect of insulin on oxidative stress. In a recent study28 assessing subjects with T1DM and T2DM receiving various hypoglycaemic treatment regimens (OAD only, OAD with insulin and insulin only), it was found that patients being treated with insulin, alone or in combination with OAD, had a lower urinary excretion rate of 8-iso-PGF2α than patients treated with OAD only.

Our 8-iso-PGF2α urinary excretion results differ from those published by other authors, on which we have commented above, as they describe lower urinary 8-iso-PGF2α values in adult populations with T1DM or T2DM.25–28 However, there are fewer studies assessing oxidative stress and its relationship with T1DM in child populations. Gleisner et al.29 described an 8-iso-PGF2α urinary excretion rate of 1672±1706pg/mg creatinine, measured by enzyme immunoassay, in a group of children with T1DM of less than five years’ duration. However, they did not find statistically significant differences between the urinary excretion rate of 8-iso-PGF2α in these children with T1DM and 13 healthy controls paired for age and sex. In their study on 48 children and adolescents with T1DM, Schreiver et al.20 found a creatinine-normalized 8-iso-PGF2α urinary excretion rate of 2530±240pg/mg creatinine, measured by tandem mass spectrometry, and were unable to establish a correlation between glycaemic variability parameters and urinary excretion rate of 8-iso-PGF2α. The urinary 8-iso-PGF2α data we describe in our study are also lower than those found in these two studies on paediatric populations. In this case, although all the study subjects have T1DM and are being treated exclusively with insulin, the methodology used to measure urinary 8-iso-PGF2α is also different (enzyme immunoassay in our study and that of Gleisner vs tandem mass spectrometry in Schreiver's study), and other factors, such as the duration of the diabetes and the type of treatment (MDII vs CSII), are also different in this study and those referenced above, which could affect the results. In any case, further studies are needed, on the one hand to assess urinary 8-iso-PGF2α values in a healthy paediatric population in order to obtain reference values, and on the other, in homogeneous groups containing a larger number of children with T1DM.

The limitations of our study include its cross-sectional design and the high percentage of participants who did not take the urine sample, which means that the relationship between glycaemic variability and oxidative stress could not be properly assessed. The glycaemic variability parameters were estimated from the data for determinations of blood glucose and not from a CGM system; however, a recent study showed that the MODD index calculated from four blood glucose measurements in a group of children with T1DM correlated well (r=0.87) with the MODD index calculated from information collected by a CGM system, and concluded that blood glucose determinations could also be used to calculate the MODD index.19

In conclusion, high intraday and day-to-day glycaemic variability and moderately high urinary 8-iso-PGF2α excretion were observed in children and adolescents with T1DM attending a summer camp. However, we did not find correlations between markers of glycaemic variability and oxidative stress measured by the rate of urinary excretion of 8-iso-PGF2α. Further studies are needed to evaluate oxidative stress and its relationship with glycaemic variability in children with T1DM.

The authors have no conflicts of interest to declare.