What About ‘Hunger Diabetes’?
Are we forgetting “Hunger Diabetes”?
In an era of personalized medicine, and specifically nutrition, putting vast sets of patient data into context of nuanced and value based care is becoming increasingly complex. In the last few years there has been fantastic progress in the field of personalized nutrition largely spearheaded by Prof Tim Spector and the founders at ZOE, towards helping people understand their bio-individuality, that is how individualized our gut, blood sugar (glucose) and blood fat response to meals is. Their nutrition science research, largest in the world of its kind, demonstrates that even identical twins can have divergent responses to the same meal. This research culminated in their seminal 2020 paper published in Nature, and subsequently in the roll-out of their at-home ZOE test kit to the consumer market.
One of the most notable features of the ZOE test kit is the continuous glucose monitor (CGM). For those who have already tested this pioneering product, you’ll know that by the simple application of a plastic disc to the arm, a small flexible filament sits just into the skin and connects with a mobile phone app. From this app, users gain a near-real time insight into the glucose response to every meal, snack or drink they ingest over the 2 weeks that each disc lasts.
Aside from the pure interest and intrigue as to how your body responds to different foods - not to mention the challenges and experiments to explore - many aspire to derive the emerging potential health benefits of having a greater insight into the metabolic impact of different patterns of eating and adjusting their lifestyles accordingly. Research has shown that even in individuals considered healthy by traditional standards, CGM data can demonstrate that a significant proportion of those testing have glucose levels reaching prediabetic and diabetic levels.
Part of the pull of this project is feeling like you are a part of a community in trying to make behavioral changes through self-accountability. The reports that you get from the kit compare your blood sugar responses to those who have previously experienced the product and submitted their figures to the ever-growing database that drives the continuously evolving algorithm.
Of course, there is always nuance in personalized care; this is why population based medicine is good for populations but not always the individual. One such nuance we’ve encountered in some of our patients is the unexpectedly unfavorable blood sugar control result from the ZOE report despite previous analysis suggesting healthy metabolic and CGM profiles. In several interviews Prof Tim Spector has noted how conventionally ‘healthy’ food like oats, bananas and mashed sweet potatoes have led to unbelievable sugar spikes in his own reports. One plausible explanation of these findings may be a reflection of several patients' adaptation to a low carbohydrate diet.
Any significant change in macronutrient intake impacts the myriad of biochemical pathways that relate to energy balance. Adaptations to transitioning to a low carbohydrate diet from a high carbohydrate diet, and vice versa, may take several weeks to develop, as demonstrated in a recent study. It has been known for decades that when an individual who typically ingests a low carbohydrate diet, suddenly ingests a glucose bolus, such as with the oral glucose tolerance diabetes test, there is an exaggerated rise in glucose and the possibility of a false positive diagnosis of diabetes. “Hunger diabetes” was first noted in the 18th Century when starved dogs that were given a high carbohydrate meal were found to have sugar in their urine.
When a physician orders an oral glucose tolerance test, a requirement should be to ingest sufficient carbohydrate in the days and even right up until the last meal before the fast prior to the test (>150g per day for 3 days and >50g per meal). Failure to do so will see as much as a 2.5-fold rise or “spike” in glucose 2 hours after, for example a muffin, and thus classification as an individual with poor glucose control when simply transitioning to a higher carbohydrate diet in advance of the test will give far more favorable results. This is a good example of how even AI based personalized recommendations are nuanced and need to be interpreted in context.
A secondary implication (but as yet, of uncertain significance) of this phenomenon is for those on a low carbohydrate diet who switch suddenly back to a high carbohydrate diet. It might be a consideration particularly for those who enjoy a splurge on carbohydrates at the weekend, when the pancreas’s beta cells are still in slumber and haven’t had a chance to wake up to release enough insulin to better control the post-meal glucose surge.
As always, context is key. These studies, no matter how well done, are based on population averages. Individuals within these nutritional studies may have equal and opposite responses, as seen in our identical twins from the ZOE studies. There is no substitute for knowing your own data and response through testing and experimentation or delegating this process to a qualified team to provide guidance and advice.