Low-energy total diet replacement intervention in patients with type 2 diabetes mellitus and obesity treated with insulin: a randomized trial
"At randomization, participants commenced a 12-week TDR [total diet replacement] formula LED [low energy diet]... followed by 12 weeks of structured food reintroduction and then ongoing followup in combination with an energy deficit diet at 3-month intervals until 12 months. For the first 12 weeks, all meals were replaced with four formula LED products per day (800–820 kcal/day, 57%
carbohydrate, 14% fat, 26% protein and 3% fiber) in addition to at least 2.25 liters of energy-free beverages. A fiber supplement was recommended, if required, to avoid constipation, a common side effect of using a TDR".
For three months patients were starved on 800kcal per day. At 56% carbohydrate that makes carbs come out at around 100g/d. Oddly enough, restricting carbs to this level allowed a drop in insulin usage. Indeed, there was such a marked drop in insulin usage that some patients coming off insulin all together. I wonder what these starvation subjects would think if you told them that they could have had equal reductions in insulin usage just by restricting the carbohydrate content of their diets to that 100g/d, while still allowing fat and protein to satiety... I suspect that a) no one has told them this and b) they might not be best pleased to find out retrospectively.
For a second three months a little food was added to their diet, but not much. For the final six months patients were kept a little hungry but not so much as in the first six months of the study.
Here is what the abstract says:
"Results: Mean weight loss at 12 months was 9.8 kg (SD 4.9) in the intervention and 5.6 kg (SD 6.1) in the control group (adjusted mean difference −4.3 kg, 95% CI −6.3 to 2.3, p less than 0.001)".
Here is what the results show for the intervention group:
Here is the same graph but simplified in to three red lines representing the three phases of the study:
You can argue the exact slopes of the lines but overall the pattern is correct. Something like this:
Now it is time to look into the future. Usually this is difficult but I think that in this case the general shape of the graph lets us predict the shape of things to come when related to weight gain. Plus, because it becomes obvious in the later months of the study (from HbA1c values) that insulin is going to have to be added back in, at this time the rate of weight gain might actually increase (dramatically), but we can't know that.
Using a simple maintenance of the status quo (best case scenario) we get this, looking forwards to around about the 24 month mark:
Weight gain, in the aftermath of a year of hunger, might not stop at baseline mass either.
I think it is also possible to look in to the future of glycaemia too, by extending the plot of HbA1c with time, working from the published graph in the results. Taken forwards to 16 months or so, it looks something like this:
Maybe I'm being pessimistic. Maybe sudden tolerance of chronic hunger might kick in and reverse the adverse trends in weight and glycaemia clearly present at the end of the study. Maybe subjects might suddenly become slim and euglycaemic.