The evidence behind dietary and lifestyle interventions for type 2 diabetes

Dietary management is widely recommended for people diagnosed with type 2 diabetes. Diabetics are referred for disease case management, often mandated by insurers, to make sure they follow their prescribed diabetes diet. Given the potential serious outcomes of the disease and the widespread claims of a grave epidemic of type 2 diabetes, you might expect that the best diet* has been established. But there are many different ideas on the dietary and lifestyle interventions that people with type 2 diabetes should follow in order to reduce vascular complications and mortality. Which one(s) are supported by clinical evidence?

Cochrane Collaborative, the most internationally respected source of objective reviews of clinical research, did a systematic review of the evidence on the efficacy of various dietary and behavioral interventions for adult type 2 diabetes and recently released their important update. Yet, it received no media attention. Given the importance of this research, you deserve to learn its findings.

Understanding when evidence is reliable and of high quality — the concept of 'fair test'

Sound, evidence-based medical care is not based on anecdotes, correlations, or the popularity of a belief. The only way to really know if a dietary or other medical intervention is safe and effective, and better than nothing (a placebo) or standard care, is a randomized, controlled clinical trial. But medical professionals know that it’s also very easy to design and conduct a study that seems to prove just about anything someone might wish to conclude. Only trials that are of good quality and unbiased — fair tests of an intervention — provide credible evidence.

In looking at quality research evidence, health benefits are also measured by actual clinical endpoints, such as reductions in rates of death and diabetes complications seen over the long-term, not surrogate endpoints (like lab tests, weight, and other health indices), subjective impressions, or short-term changes in indices. It makes no difference if lab numbers meet some arbitrary threshold, for example, if patients are more likely to suffer complications and die prematurely. And if an intervention works for 0.1% of patients but not for the other 99.9%, or harms more than it helps, it would also [ideally] not become a clinical care guideline prescribed for all patients.

As the Cochrane reviewers noted in their introduction, a diet is usually the first treatment prescribed when someone has been diagnosed with type 2 diabetes. And there are many different types of diabetes diets being proposed, each one with its vehement supporters: low-fat and high unrefined carbohydrates (high fiber, whole grain), low-carb, low glycemic index diets, calorie-reduced diets, and various combinations.

“Recently, research has also looked into the effect of individual macro- and micro-nutrients such as fibre, chromium, and so-called ’functional foods’,” they said. But there is little evidence to support these, they said, nor long-term intervention trials conducted to evaluate them. Not included in this review were studies specifically evaluating the effects of fish oils (omega-3) on type 2 diabetes, however, as another recent Cochrane review had already examined this and found no evidence for a statistically significant effect on glycemic control or fasting insulin, nor had studies shown benefits to vascular endpoints or mortality.

The role and type of dietary advice following diagnosis of type 2 diabetes, and its efficacy during long-term follow-up, is surprisingly unclear, they noted. “While short-term benefits are assumed, attempting to treat type 2 diabetes using diet alone is not a particularly successful long-term intervention, as illustrated by the United Kingdom Prospective Diabetes Study,” they wrote. Despite reinforced dietary interventions, after three years only 20%, and after nine years only 8%, of type 2 diabetics maintained fasting blood glucoses of less than 7.8 mmol/L. Medications become necessary when diet and exercise fail to control blood glucose to currently recommended levels, they said. But, it’s not known if dietary interventions change the progression of the disease process itself.

The studies

In 2007, the Cochrane reviewers had conducted a comprehensive systematic review and QUOROM (evaluation of the quality of the studies) of 8,675 studies, after searching for all of the randomized, controlled clinical trials of dietary advice and interventions in type 2 diabetes that had lasted at least six months. It had found no evidence to support dietary interventions. This just-published update included background research and a review of 1,413 additional studies that had been excluded in 2007 to determine if any type of dietary advice or weight loss, with or without other lifestyle interventions, could be shown to improve morbidity, total mortality, or quality of life.

Eighteen clinical trials of 1,467 type 2 diabetics met the inclusion criteria. Study participants were representative of the general population of type 2 diabetics. “Nine studies focused on looking at the effects of two types of diabetic dietary advice that did not differ in intent to lose weight, three studies focused on looking at dietary advice versus dietary advice plus behavioural approaches, and six studies concentrated on dietary advice versus dietary advice plus exercise.”

As with all Cochrane systematic reviews, the quality of each trial was assessed, using those established determinants of bias and fair test. So, for instance, minimization of selection bias (randomization method), detection bias (blinded), and attrition bias (dropout rates) were carefully evaluated.

Findings

Each trial differed in comparing special dietary interventions to regular diets or to various other dietary interventions, but the reviewers came to similar conclusions with each one. As they described:

· The two studies comparing the American Diabetes Association exchange diet to a regular low-fat diet were found to be highly biased, with no micro and macrovascular diabetic complications or glycated hemoglobin changes reported.

· The five studies assessing low-fat versus moderate-fat or low-carb diets were also found to be at high risk of bias, with micro and macrovascular diabetic complications only reported in one, in vague terms, and not at all in the others. No conclusions could be drawn form the marginal, untenable changes in glycated hemoglobin levels reported, they said.

· Two studies of very low-calorie diets (400-500 kcal/day) versus a traditional low-calorie diet (1,00-1,200 kcal/day) were both found at high risk for bias. Glycated hemoglobin levels dropping in accordance to the initial weight loss, but rebounded by 24 months back to baseline. Neither reported vascular complications and no conclusions of a benefit in calorie reductions could be made, they said.

· Six trials compared dietary advice to dietary advice and exercise. The quality of the trials, however, suffered high bias. Exercise along with weight loss was suggested as being associated with a 1% reduction in glycated hemoglobin levels at 12 months, but the high degree of bias should be considered before making interpretations, they wrote. None of these studies had reported vascular complications.

· Three studies assessed dietary advice versus dietary advice plus behavioral interventions and all three trials had high risk of bias and “the development of micro and macrovascular diabetic complications was not reported in any of the studies.” Glycated hemoglobin was reported in two of the trials, but there was more improvement in glycemic control seen in the usual care group compared to the intense behavioral interventions, although the changes were not significant.

Reading their detailed analyses of all of the dietary intervention clinical trials to date in their 73-page report revealed a stunning lack of evidence. Trials to date have all been short-term, examined surrogate endpoints and been unable to demonstrate benefits in actual clinical outcomes. Eating a variety of foods for good nutrition is one thing, but no specific dietary regimen has been able to show an improved outcome for type 2 diabetics. They said in their summary that:

Only a minority of the trials examined hard clinical endpoints (such as death or development of macrovascular or microvascular diabetic complications), and those that did offered no details; most articles concerned themselves with the reporting and discussion of the participants’ weight and blood glucose control.

Weight loss is most popularly believed to be an important part of type 2 diabetes management, but no diet has been shown effective. As they said, another recent review published in the Cochrane Library had found no real differences in weight loss between low-fat diets and other weight-reducing diets, all of which were “clinically insignificant.” In this review, they said, “clinically meaningful differences in glucose profile were not achieved” in trials of “low-fat or other weight-reducing diets.”

In conclusion:

There were not enough data in the studies that assessed one dietary advice versus another (different) type of dietary advice to enable us to reach any satisfactory conclusions. The data included in the trials in this review which assessed dietary advice plus behavioural approaches did not have the data to allow us to reach any satisfactory substantial conclusions.

Despite all of the claims for the benefits of various dietary and lifestyle interventions for type 2 diabetes, none of them are yet supportable by quality evidence. The Cochrane reviewers expressed concern about the lack of hard, clinical endpoints in research to date and said there is a need for future researchers to “take care to record and publish mortality data, changes or delays in medication needs and quality of life, as these are the outcomes of importance to people with type 2 diabetes.”

© 2008 Sandy Szwarc

* Diet in this context refers to an eating plan, not specifically a weight loss diet.

Thank you Morley!