The effect of carbohydrate intake on glycaemic control in individuals with type 1 diabetes: a randomised, open-label, crossover trial

Summary

Background

Few studies have examined the effects of lower carbohydrate diets on glucose control in persons with type 1 diabetes (T1D). The objective of the study was to investigate whether a moderate carbohydrate diet improves glucose control in persons with T1D.

Methods

A randomised, multicentre, open-label, crossover trial over 12 weeks.

There were 69 individuals assessed for eligibility, 54 adults with T1D and HbA1c ≥ 58 mmol/mol (7.5%) were randomised.

Interventions were moderate carbohydrate diet versus traditional diet (30 vs 50% of total energy from carbohydrates) over four weeks, with a four-week wash-out period between treatments.

Masked continuous glucose monitoring was used to evaluate effects on glucose control. The primary endpoint was the difference in mean glucose levels between the last 14 days of each diet phase.

Findings

50 individuals were included in the full analysis set with a mean baseline HbA1c of 69 mmol/mol (8.4%), BMI 29 kg/m², age of 48 years, and 50% were female.

The difference in mean glucose levels between moderate carbohydrate and traditional diet was −0.6 mmol/L, 95% CI −0.9 to −0.3, p < 0.001. Time in range increased during moderate carbohydrate diet by 4.7% (68 min/24 h) (95% CI 1.3 to 8.0), p = 0.008. Time above range (>10 mmol/L) decreased by 5.9% (85 min/24 h), 95% CI −9.6 to −2.2, p = 0.003.

There were no significant differences in the standard deviation of glucose levels (95% CI −0.3 to 0.0 mmol/L, p = 0.15) or hypoglycaemia in the range <3.9 mmol/L (95% CI −0.4 to 2.9%, p = 0.13) and <3.0 mmol/L (95% CI −0.4 to 1.6%, p = 0.26).

Four participants withdrew, none because of adverse events. There were no serious adverse events including severe hypoglycaemia and ketoacidosis. Mean ketone levels were 0.17 (SD 0.14) mmol/L during traditional and 0.18 (SD 0.13) mmol/L during moderate carbohydrate diet (p = 0.02).

Interpretation

A moderate carbohydrate diet is associated with decreases in mean glucose levels and time above range and increases in time in range without increased risk of hypoglycaemia or ketoacidosis compared with a traditional diet in individuals with T1D.

• Evidence before this study

We searched PubMed with end date on May 15, 2017, using the search terms “carbohydrate restriction”, “low carbohydrate diet”, “LCD” and “type 1 diabetes” in combinations.

The titles and/or abstracts were screened and selected manually. Very few studies were found although the ones found were indicating positive effects on glucose control. Concerns about elevated risk of dyslipidaemia, hypoglycaemia and diabetic ketoacidosis also existed.

The studies found were either observational, small, not randomised, or lacked control groups. An update search in PubMed on May 15, 2023, using the same search terms did not reveal any new relevant evidence regarding glucose control or safety from larger randomised studies.

• Added value of this study

In this randomised, crossover trial over 12 weeks including 54 adults with type 1 diabetes the primary outcome mean glucose level was significantly reduced by −0.6 mmol/L (−11 mg/dL) with moderate restricted carbohydrate diet compared to traditional diet.

Other endpoints showed more time in range (3.9–10.0 mmol/L), time in tighter ranges (3.9–7.8 and 3.5–7.8 mmol/L), and less time with high and very high glucose levels as well as increased treatment satisfaction.

There was no increase in time in hypoglycaemia or cardiovascular risk factors such as lipids and blood pressure, no ketoacidosis, severe elevated ketone levels, or serious adverse events during the trial.

• Implications of all the available evidence

This study shows that a moderate carbohydrate diet is more effective than a traditional diet with a higher amount of carbohydrates in terms of decreasing mean glucose levels and time above range and increasing time in range and treatment satisfaction without increased risk of hypoglycaemia, dyslipidaemia, or ketoacidosis in individuals with type 1 diabetes.

These results show that a healthy moderate carbohydrate diet can be considered as a safe and effective treatment option for individuals with type 1 diabetes which extends possibilities for more differentiated diabetes care and provide further options to individualising diet treatment.

From the article Discussion

A moderate carbohydrate diet leads to significantly improved glucose control compared to a traditional diet with higher carbohydrate content in adults with T1D. The primary outcome, mean glucose level, was significantly reduced. Other endpoints showed more time in range, time in tighter ranges, and less time with high and very high glucose levels as well as increased treatment satisfaction. There was no increase in time in hypoglycaemia or cardiovascular risk factors such as lipids and blood pressure, no ketoacidosis, severe elevated ketone levels, or adverse events during the moderate carbohydrate diet.

The lack of diet studies in T1D may be due to need for them to be driven academically, little interest from the pharmaceutical industry, and their time-consuming nature. A few earlier studies showed improvements in glucose control with low carbohydrate diets in adults with T1D, but there have also been concerns about safety regarding the risk of elevated lipids and increased risk of ketoacidosis and hypoglycaemia.

Because evidence regarding efficacy and safety are lacking and the possible increased risk of ketoacidosis and severe hypoglycaemia, low carbohydrate diets are not yet recommended in dietary guidelines for people with T1D.

Earlier studies were either not randomised or very small. Low carbohydrate diets can differ widely in recommended carbohydrate amount, and most studies were of very low carbohydrate intake (less than 120 g and 5% of energy).

In this study our goal was to ensure a healthy diet to increase adherence for longer time periods. Safety was also a priority; therefore, we used a moderate carbohydrate diet with 30% of energy from carbohydrates compared to the traditional 50% which would make a clinically relevant difference in carbohydrate amount, but still likely be safe in both the short- and long-term. It also included healthy foods such as wholegrains, legumes, vegetables, and unsaturated fats according to guidelines to avoid negative health effects such as elevated blood pressure or dyslipidaemia.

The obtained effect on mean glucose corresponds to approximately 3 mmol/mol (0.3%) reduction in HbA1cwhich has been associated with a reduced risk of retinopathy.

Mean glucose level was chosen as the primary endpoint since the true glucose level without increased time in hypoglycaemia is a marker of both acute and long-term complications. Moreover, an increase in time in range of 5%, comparable to the current results of 4.7%, is considered a clinically relevant difference according to CGM guidelines.

Although participants increased in time in range the majority did not reach targets of 70 percent TIR indicating the need of additional improvements in diabetes care for many persons to achieve targets of glucose control. Further positive effects were seen in the decrease in time above range (>10 mmol/L) by 5.9% as well as the decrease in time above range (>13.9 mmol/L) of 3.6%. Since an exponential relationship exists between glucose levels and diabetes complications it is important to reduce time with very high glucose levels for all individuals with T1D.

One possible mechanism for the decrease in mean glucose level as well as the other positive effects on glucose control during the moderate carbohydrate diet may be due to fewer glucose excursions and peaks. Predicting the precise amount of insulin required becomes challenging due to several factors influencing insulin uptake such as physical activity during the day. Consequently, reducing the carbohydrate intake per meal potentially mitigates glycaemic peaks even if the insulin dosage is not entirely correct.

The current results indicate that a moderate carbohydrate diet should be included in dietary guidelines for persons with T1D as an alternative for decreasing mean glucose levels and that it can be considered safe, although it is important that it is possible for individuals to receive dietary advice from a dietitian to make sure the diet is healthy in terms of fat and carbohydrate sources.

Strengths of the study include randomised crossover design which reduces risk of confounding factors and person-to-person variation. It also included a detailed individual diet plan for each participant and during each diet phase with regular follow-up to increase adherence. Finally, it included a structured insulin management plan to keep it as comparable as possible between diet phases.

In this study participants measured ketone levels twice a week in the morning and evening, and they were asked to report adverse events during follow-up to determine risk of ketoacidosis and other possible negative effects of the diet. During moderate carbohydrate diet, ketone levels were slightly elevated, although levels overall were low and never were severely elevated. There were no cases of ketoacidosis, indicating that this level of carbohydrate intake may be safe. Insulin to body weight ratio was slightly decreased in the moderate carbohydrate diet, indicating that insulin doses were decreasing along with carbohydrate intake, which logically would be expected.

The total study period of 12 weeks and diet phases of 4 weeks each was intentional making it possible for participants to comply with dietary changes and all study procedures including keeping food diaries, measuring ketone levels, and using masked CGM to provide detailed glucose and diet data. This study mainly elucidates physiological effects and effects on glucose control of a moderate carbohydrate diet, and individual preferences likely exist which should be considered in clinical practice in order to increase compliance during longer treatment periods.

The food diaries indicated that differences in carbohydrate intake between the phases were less than planned, and a minority of participants lacked food records. Furthermore, carbohydrate intake was self-reported and thus may be biased.

Although food diaries were not registered in both phases by all participants the majority had recordings and the per protocol analyses using these data confirmed effects on glucose control for individuals differing in carbohydrate intake. Of note is that only adults with HbA1c of ≥58 mmol/mol (7.5%) were included in the study which may restrict results to this population.

Although the primary endpoint, mean glucose level, can be viewed as confirmatory, other glucometrics should be viewed as exploratory and preferentially confirmed in other studies. However, since there is a strong correlation between mean glucose level, time in range, and time in hyperglycaemia, our findings further support likely beneficial effect on other important glucometrics.

In summary, this study shows that in persons with T1D a moderate carbohydrate diet is more effective than a traditional diet with a higher amount of carbohydrates in terms of decreasing mean glucose levels and time above range and increasing time in range and treatment satisfaction without increased risk of hypoglycaemia, dyslipidaemia, or ketoacidosis.

These results show that a healthy moderate carbohydrate diet can be considered as a safe and effective treatment option for individuals with T1D which extends possibilities for more differentiated diabetes care and provides further options to individualising diet treatment.

https://www.thelancet.com/action/showPdf?pii=S2666-7762%2823%2900218-1

Nyhetsinfo

www red DiabetologNytt