October 26, 2022

expert reaction to study looking at effects of a low-carbohydrate dietary intervention and improved glycaemia

A study published in JAMA Network Open looks at effects of a low-carbohydrate dietary intervention on haemoglobin.

Professor Nita Forouhi, MRC Epidemiology Unit, University of Cambridge, said:

“This research confirms previous findings from many clinical trials and meta-analyses of clinical trials that, among people with type 2 diabetes, a low-carbohydrate diet can improve glucose control to a greater degree than with a usual diet, but this is limited to being over the short term up-to six months. In this regard the findings are not novel but confirmatory. A critical factor is to know if this dietary approach is sustainable and effective in the longer term, but this current study cannot address that.

“The things that are positively different about this study are that it includes people with prediabetes as well as previously unknown type 2 diabetes, a greater diversity of participants with African-Americans as well as White Americans, and that in exploratory analyses it used continuous glucose monitoring to look at variability in glucose levels dynamically for 14 days in a sub-set of participants and found directionally similar results. The dietary intervention in this study was limited to a healthy low-carbohydrate diet rather than any low-carbohydrate diet, which is an important public health consideration. 

“The findings cannot be considered conclusive and the authors themselves call for further research because interpretation is limited by many factors. In particular: it is possible that the observed effects may not be solely ascribed to a change in diet but may be in part due to intensive behavioural support in the intervention group with the frequent individual and group support as well as the provision of key supplemental food. The magnitude of effect on the improvement in HbA1c is modest in the main analysis (a net difference of 0.23% lower HbA1c in the intervention compared with comparator group), and even smaller when statistical adjustments are made (a net difference of 0.17%). It is unclear if effects on glucose would have been observed without weight loss or independent of calorie intake, which begs the question of whether a low calorie diet would be just as or even more effective, and for which there is evidence from other studies such as the UK-based DiRECT study. The reported calorie intakes at baseline seem quite low at around 1890 kcal/day and 1789 kcal/day in the low-carbohydrate and usual diet groups with unless the participants were already on a weight reducing diet. This is likely under-reporting when considering that the mean BMI of the participants was high around 35 kg/m2. It is notable that 34% of participants reported cramps in the intervention group vs. 19% in the usual diet group at six months and it is unclear if this side effect may impact adherence in the longer term and when free-living outside of the setting of a clinical trial which included participant support.

“Overall, this study adds to the growing literature that a healthy low-carbohydrate diet can meaningfully be part of the management of type 2 diabetes and now with evidence also for prediabetes, but longer term effects are unknown and current limitations mean this is not definitive, other dietary approaches may also be equally or more effective, and no one size may fit all.”

Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:

“This is an interesting study, which compared a low carbohydrate diet which the authors described as healthy against the usual dietary advice for people with relatively well controlled type 2 diabetes who were not taking any medication for their diabetes. It took 75 people and put them on a low carbohydrate diet, aiming for around 40g of carbohydrate per day along with weekly follow up for a month followed by regular group sessions and telephone calls against 75 people who were just given information and the option of attending some non-food related groups. This is the first main source of confounding, in that the groups were not given the same level of input, so the group of the low carbohydrate groups who were enrolled on a courses called Go Low and Keep it Low were given a lot more support than the control group. This could be to support the effect of the education programme, ideally both groups should have been given the same level of support.

“The second challenge is it is not possible to tell if the low carbohydrate element was responsible for the significant reduction in HbA1c in the low carbohydrate group or the fact that they as a group were heavier and lost more weight. Although low carbohydrate diets, if well designed and healthy are acknowledged as a great approach to manage type 2 diabetes and even bring it into remission is it not clear if this is because they prevent rises in blood glucose (sugar) or if they help people lose weight, which helps people to be more insulin sensitive and allows their muscles and liver to use glucose more efficiently. So, this could show that the regular contact as well as their weight regressing to the mean could explain the effect not just the low carbohydrate diet.

“Although low carbohydrate diets have been shown to be very effective at managing diabetes over 3-6 months as seen in this study, which has also been shown in general practice in the UK to be able bring glucose and HbA1c (a long term measure of glucose over 3 months) to levels which are considered to have brought type 2 diabetes into remission (below 6.5% or 48mmol/mol for more than 6 months without diabetes medication). As many of the participants in this study in both arms were already below this target (which is also used to diagnose diabetes), although more of the low carbohydrate arm achieved a HbA1c below 6.0% (42mmol/mol) which might be considered to suggest they no longer had pre-diabetes – as they all started with good control of their diabetes, it possibly accounts for the significant but modest change in the low carbohydrate diet group compared to many studies of aiming to show an improvement in diabetes control.

“So, overall this adds to the many studies showing that low carbohydrate diets can be effective over a period of 6 months, adding to earlier studies from Australia which have shown it can be effective for up to 2 years. This perhaps adds that it can help even those people with type 2 diabetes who are well controlled to manage their diabetes and lose weight. It is good that this study recorded not only the 40g per day diet recommended but also tried (although limited by how accurately people self-reported their own diet) the amount of carbohydrate they were eating as the study went on, and although they did not achieve the target of initially 40g per day rising to 60g per day, a level of carbohydrate restriction can be useful to support management of type 2 diabetes. If someone is thinking of changing their diet to help manage their type 2 diabetes, including following a low carbohydrate diet it is important they talk to their healthcare team and ideally a dietitian. As if they take medication for their diabetes, they may need advice on how to adjust their medication before starting such a diet.”

Prof Keith Frayn, Emeritus Professor of Human Metabolism, Radcliffe Department of Medicine, University of Oxford, said:

“This study purports to show that a low-carbohydrate diet improves diabetes control as measured by HbA1c.  However, what it really shows is that a calorie-restricted diet is effective in this respect.  It was not a comparison of a low-carbohydrate diet with any other calorie-reduced diet.  We already know that reducing calorie intake will improve diabetes control.  And we also know that, whilst low-carbohydrate diets can help people to lose weight over a short period, such as 6 months, as in this study, they bear no advantage over other calorie-restricted diets in the longer term.  This study is best seen as confirmation that a reduced calorie diet is beneficial to diabetes control, rather than telling us anything new about low-carbohydrate diets in particular.”

Prof Naveed Sattar, Professor of Metabolic Medicine, University of Glasgow, said:

“There is nothing new here – we know weight loss leads to improvements in glucose – and so a low carb diet, known to lower weight, will predictably lower HbA1c – what this study cannot show due to design is to what extent benefits are due to weight loss (likely most in my opinion) versus cutting carbs per se.  That is the bigger question though low carb diets may not be as good for people in terms of cardiovascular risk and may not be sustainable over longer periods.”

‘Effects of a Low-Carbohydrate Dietary Intervention on Hemoglobin A1c: A Randomized Clinical Trial’ by Kirsten S. Dorans et al. was published in JAMA Network Open at 16:00 UK TIME Wednesday 26 October 2022.

DOI: 10.1001/jamanetworkopen.2022.38645

Declared interests

Professor Nita Forouhi: “No conflicts of interest to declare.”

Dr Duane Mellor: “No conflicts of interest.”

Prof Keith Frayn: “I have no conflict of interest to declare.”

Prof Naveed Sattar was co-investigator in the DIRECT trial and has consulted for several companies that market glucose lowering agents.

For all other experts, no response to our request for declarations of interest was received.