Assessment of dietary intake by self-report in adult patients with type 1 diabetes treated with a personal insulin pump

One of the most important factors to affect general health is appropriate nutrition [1, 2]. Medical nutrition therapy is of paramount importance in the modern treatment of diabetes [3, 4]. Dietary management approaches that are tailored to meet the needs of people with type 1 diabetes must reflect recommendations aimed at reducing the risk of acute and chronic complications. Unfortunately, dietary intake in young populations with type 1 diabetes often does not follow dietary guidelines [5]. These patients should pay attention not only to the intake of carbohydrates but also to the nutritional quality of the meals. The exaggerated focus on planning meals only in relation to carbohydrate intake can pose a risk to an otherwise generally healthy diet [6]. Available research seems to indicate that diets focusing on higher carbohydrate [7, 8], lower fat [7, 8, 9], higher fiber [9], and higher fruit and vegetable [10, 11] intake have a positive influence on glycemic control. On the other hand, some studies show no such association [12, 13, 14], and several studies suggest the opposite – that lower carbohydrate consumption [15] produces better glycemic control. As far as rational nutrition is concerned, there is hardly any difference between the recommendations for individuals with type 1 diabetes, and the general population of healthy people [3, 4].

Diabetes-related malnutrition can be caused by poor dietary intake or malabsorption, which can lead to protein deficiency and other nutrients, affecting energy levels as well as several organs and systems, and being responsible for alterations in the digestive, immune, and muscular functions [16]. Contemporary advances in diabetes management, such as personal insulin pumps, allowed for lifestyle flexibility for people with T1DM to such an extent that the individuals can be tempted not to follow the prevailing guidelines for a healthy diet. This can prove especially important in terms of its effect on the development of long-term macrovascular complications of the disease [17, 18].

Our study aimed to assess the implementation of dietary recommendations and their relation to metabolic control in a group of young adults with T1DM treated with personal insulin pumps.

Our results include data from 48 patients: 15 females (31.2%) and 33 males (68.8%) with type 1 diabetes with a median age of 25.6 years [22.2–28.3] and 15 years of diabetes duration (± 6.6). Most of the patients have their BMI within a normal range 32 (66.7%), 13 (27.1%) were overweight, 2 (4.2%) were obese and 1 (2.1%) was underweight. All the participants were Caucasian. Detailed patient characteristics are presented in Table 1.

In the present study, we have indicated that our cohort of patients with type 1 diabetes is characterized by insufficient consumption below EAR/AI of most nutrients and vitamins: sodium, potassium, calcium, magnesium, iron, zinc, copper, iodine, manganese, vitamin A, vitamin D, vitamin E, thiamine, riboflavin, niacin, vitamin B6, folate, vitamin B12, vitamin C. There was no significant difference for consumption of most nutrients regarding metabolic control (except PUFA, niacin, sodium, calcium).

The analyzed subgroups divided based on glycated hemoglobin A1C, did not differ significantly in baseline characteristics, except for BMI (22.7 ± 2.4 [HbA1c<6.5%] vs. 25.6 ± 6.2 [HbA1c≥6.5%], p<0.04). All the patients were free from advanced late complications of diabetes.

Mean total energy per day was 2012 ± 684 calories, there was no significant difference in subgroups. The daily supply of calories depends on various factors including gender, age, weight, and physical activity level (PAL). Observed value is appropriate for a man with low PAL or a woman with medium PAL. Most of our patients obtained fewer calories than recommended for total energy expenditure [TEE]. In the literature there is a strong suspicion of non-reporting all meals or inaccurate weighing: both lead to underestimation of calories (observed level of underestimation is between 6% and 40%) [20, 21, 22]. A recent meta-analysis has confirmed a significant underestimation of total energy intake [TEI] in population samples of adults when energy intake was estimated by various retrospective and prospective dietary assessment methods (24h diet recalls, estimated food records, weighed food records, and diet histories) in comparison to an objective reference measure of TEI using doubly labeled water. Using 24h diet recall, men underestimate total caloric in-take by a mean of 715 kcal and women by 633 kcal per day. No significant differences in underestimation were identified based on sex (except estimated food records, for which males underestimated energy intake more than females) [23]. As early as 1990 it was shown that underreporting increased with increased BMI [24]. Furthermore, according to the literature, it is acceptable to use a low-calorie diet in type 1 diabetic patients with low physical activity and accompanying overweight or obesity (which was the case in one-third of our cohort) [25, 26].

Based on our results, we may conclude that observed calorie deficiency can be partly attributed to underestimation due to the self-reporting method used but according to the literature, observed underestimation falls within the error range.

The comparison of subgroups revealed no difference in percentage intake of carbohydrates, proteins, and fats (Table 3). According to the 2021 Guidelines on the management of patients with diabetes, there is no sufficient evidence to determine the single optimal carbohydrate content in the diet. Carbohydrates should provide about 45% of the total calorie intake; if they are supplied in the form of low glycemic index and high fiber content products, their share in the total calorie intake may even be as high as 60%. A high-calorie diet is intended for people who are very physically active, in contrast to the low carbohydrate diet (25–45% in total calorie intake) which may be temporarily recommended in patients with little physical activity. In our subgroups, carbohydrates constituted 52.3% ± 5.0% vs. 50.2% ± 8.2%, p=0.32 respectively. The intake of carbohydrates was about 250 grams per day in each subgroup, therefore it can be classified as a high carbohydrate diet [27]. This type of diet is meant for subjects who are very physically active, otherwise, it could be responsible for increases in glycemic variability, time spent in hypoglycemia, and gaining weight [28]. The intake of fiber was 19.1 g ± 7.3 g vs. 19.9 g ± 7.4 g, p=0.72, respectively; these values are below the recommended 25 g per day or 15 g/1000 kcal [3]. A low-fiber diet is typically observed when using highly processed food, which is popular nowadays in Europe. The consequences, however, are less glycemic control, blood lipid disorders, and the appearance of inflammation [29]. Therefore, the addition of fiber supplements, particularly containing soluble fibers, should be considered [30]. The percentage of intake of calories from fats in the studied group was between recommended 25% and 40% in both groups. However, values of saturated fatty acids are higher than recommended (less than 10% of the total calorie intake). When the intake of both monounsaturated and polyunsaturated fats is insufficient, a diet rich in saturated fats leads to increased serum LDL cholesterol levels and cardiovascular consequences [31]. In our report, the percentage intake of polyunsaturated fatty acids was 3.9% (3.4%–4.7%) vs. 4.9% (4.0%–6.2%), p<0.04. In contrast to other studies, a higher intake of PUFA was present in a subgroup apart from the therapeutic goal [32]. Diets poor in unsaturated fatty acids are characteristic of populations in Central and Eastern Europe. One of the representatives of MUFA is oleic acid, which is olive oil, an important component of the Mediterranean diet. PUFA belong to linoleic and arachidonic acid, they are found in nuts, plant oils, and fatty fish [33]. Both unsaturated fatty acids have multidirectional action. One of the most important is the cardioprotective effect of PUFA, connected with a decrease in cardiovascular risk, which results in lower stroke and myocardial infarction rate [34]. The cardioprotective impact is due to antiarrhythmic, antithrombogenic, and anti-inflammatory effects [35]. Interestingly, PUFA could inhibit lipogenesis and increase the sense of satiety. Finally, it has a positive influence on intestinal microbiota [36]. The Guidelines indicate that protein intake should constitute about 15–20% of total calories, without details concerning sources. In our analysis, these values were within the normal limits; the reduction of intake of protein is recommended in the case of chronic kidney disease [37].

The sodium intake was above the recommended range, 126.6% (98.0%–154.4%) vs. 155.4% (116.2%–208.2%), p=0.05 (0.0458) (Table 4). High dietary sodium in T1D is common due to adding salt to highly processed food. High sodium relates to vascular dysfunction, independently of other dietary intakes, blood pressure, and glycemic control [38]. Sodium is also independently associated with all-cause mortality and end-stage renal disease [39]. Although the causality of these findings is poor, these support the opinion of exercising caution before applying salt restriction universally [39]. Calcium intake in regard to EAR was 82.7% ± 36.2% vs. 106.8% ± 44.8% p=0.05, phosphorus intake was 187.5% (155.7%–261.8%) vs. 253.5% (171.5%–302.4%) p=0.21; in both subgroups, the diet was rich in phosphates. Dietary calcium and phosphorus have an impact on bone, kidneys, and parathyroid glands; a typical modern diet, which is high in phosphorus and low in calcium, may cause secondary hyperparathyroidism and bone loss [40]. Another consequence of phosphate overload is crystal nephropathy, therefore increase of calcium in the diet or as a supplementation alleviates the detrimental effects of excess dietary phytate through excretion of undigested complexes as feces [41]. The intake of zinc in both subgroups is sufficient, without significant difference, p=0.18. The impact of zinc on glycemic control is not clear [42]; however, in the era of COVID-19, it could be important to receive the recommended dose of these micronutrients [43]. In our report, the intake of iodine is insufficient in both subgroups, 26.4% (19.8%–42.2%) vs. 35.9% (22.2%–48.5%) p=0.18. Nowadays, salt is iodized, which prevents iodine deficiency; this is especially important for pregnant women and children.

As for the vitamins, in both analyzed subgroups, the intake of vitamin D is much below the adequate intake, 12.9% (6.9%–16.7%) vs. 9.6% (6.8%–16.1%) p=0.37, respectively. Vitamin D deficiency is an important public health problem worldwide and also in Poland. It is a significant risk for both skeletal and non-skeletal disorders and several lifelong negative health outcomes [44]. Patients with type 1 diabetes are at risk of vitamin D deficiency, supplementation should be implemented and followed up under the control of 25(OH)D concentrations, in order to maintain the optimal concentration of >30–50 ng/ml [45]. In both subgroups, the intake of niacin is above estimated average requirement, 132.9% (106.9%–167.0%) vs. 184.2% (137.0%–253.7%), p=0.01. There is a lack of control trials concerning the impact of niacin on lipids and glucose control in patients with type 1 diabetes, although the effect of these vitamins in patients with type 2 diabetes is better known [46]. Analysis of the results showed that niacin alone or in combination significantly improved lipid abnormalities but requires monitoring of glucose in long-term treatment. Another inconvenient side effect is flushing [47]. The intake of folate was a little below recommendation, 88.6% ± 35.1% vs. 95.7% ± 40.0%, p=0.52; this microelement is crucial for the neural tube development of the fetus, therefore supplementation before and during the first trimester of pregnancy is highly recommended [48]. All above observed deficiencies and excesses do not cause significant clinical manifestations in our cohort. This could be indirect proof that self-reported 24h diet recalls are not full (they may be underestimated), but on the other hand, these deficiencies are observed with no clinical manifestation in similar degrees in the literature of a healthy person. For example, the Polish population tends to take higher energy from saturated fatty acids than dietary recommendations [49].

The study has several limitations. First, the sample size is rather small, and the time of observation was only 3 days. Perhaps the introduction of a mobile application, which makes it easier to count mass and calories, would encourage patients to participate in this type of valuable project [50]. Secondly, due to questionnaire research and patients’ records, there is a risk of under- or overestimation of intake. Some of the observed deficiencies can be attributed to the underestimation of total caloric intake, but this underestimation falls within the range observed in the literature. Nevertheless, 24h diet reports spanning a few days remain the most popular research tool for collecting nutritional data. Another shortcoming is that the diet database does not include information about the possible supplementation used.

The novelty of our study is the fact that there are only a few dietary studies that concern young adults with T1DM [51]. An appropriate diet could improve their quality of life and prevent irreversible complications. Finally, these data may give feedback to patients on what to increase or reduce in their everyday diet [52]. The originality of the present study is a population of young adults with type 1 diabetes, who were being treated with personal insulin pumps. This group of patients is still not widely examined [53]. Also, the strength of our study is the complexity and extensive analysis of diet.

In the population of our patients, there is an evident percentage of patients with nutrient consumption below the recommended EAR/AI. The deficiency of vitamin D and iodine may be an indication for supplementation. Our study has shown that vitamin and mineral intakes among participants with type 1 diabetes, patients with HbA1c≥6.5%, and patients with HbA1c<6.5%, were similar. Interestingly, differences between PUFA, sodium, niacin, and calcium intakes were observed only when participants were grouped with respect to glycated hemoglobin at the level of 6.5% (nonsignificant differences for HbA1c 7%). A longer observation or follow-up study is necessary to investigate the impact of diet on metabolic control and control for underestimation results.