Low GI diets and weight loss: blood sugar is only part of the story

Low GI diets can improve blood sugar management, but weight loss is a separate question. Trials show why the comparison diet and population matter.

日本語で読む（Japanese version）

Choosing lunch can turn into a blood sugar quiz. Should you keep the rice, switch the main carbohydrate, or choose whatever carries a low GI label? You started by wanting the scale to move. Now the meal seems to require predicting what will happen to your blood glucose before you can decide what to eat.

Weight loss trials give a less straightforward answer than the label suggests. An updated systematic review in people with overweight or obesity found an average difference of −0.82 kg between low GI or low GL diets and higher GI or GL diets. Yet the range of uncertainty included no difference, and the authors concluded that the diets probably produce little to no difference in weight change.[9]

That does not make low GI eating pointless. Improved blood sugar management has been reported in people with diabetes.[2] It means that blood glucose and body weight need separate scorecards, even when the results come from the same dinner table. The useful question is what happened when people changed their diets, how those diets differed, and which outcome the researchers actually measured.

What a low GI label can reasonably promise

GI stands for glycemic index. Here, it is a dietary measure concerned with the rise in blood glucose after eating. Low GI diets have been studied as a way to reduce those rises and produce more stable blood glucose concentrations.[11] GL stands for glycemic load, another measure used alongside GI to define the diets being compared.[9] The research discussed here evaluates changes in eating patterns and their outcomes. It does not provide a shopping list ranked by the amount of weight each food will make you lose.

The proposed chain of events is appealing: a smaller rise in blood glucose could mean less hunger, which could help with weight loss. But each link needs evidence. A 2018 systematic review used this hypothesis as its starting point and found a small overall improvement in body weight. It did not establish whether reduced hunger was the mechanism responsible for that improvement.[15] A plausible explanation can help formulate a research question without already answering it.

The word “low” also has a comparison built into it. Some trials compare lower and higher GI diets. Others compare a low GL diet with a low fat diet.[9][18] These are different questions. If both become the headline “low GI works,” the reader loses the information needed to understand what changed. A meal plan can differ in its carbohydrate choices while also differing in fat, fiber, or the amount eaten.

Think about what you mean when you say you are changing your staple food. Are you replacing only that part of lunch? Are the side dishes changing too? Does the new meal contain a different amount of food? Those questions do not require assigning a number to every item. They help describe the change you are actually making. In a trial that adjusts an entire dietary pattern, the result belongs to that combination of changes.[8]

A low GI label can therefore be a starting point for thinking about food, but it is an incomplete explanation of a weight result. You can pay attention to carbohydrate choices while still evaluating the whole plate. That wider view is especially useful when a diet sounds as if it has reduced a complicated decision to a single score.

Average weight differences are often small and uncertain

The 2023 systematic review brought together randomized controlled trials in people with overweight or obesity. In these trials, chance determined which dietary intervention each participant was assigned to follow. The review included 10 studies and 1,210 participants, with dietary interventions lasting from 8 weeks to 18 months. Studies of children and older adults were included, although most studies involved adults.[9]

Its weight analysis comparing low GI or GL diets with higher GI or GL diets used 7 studies and 403 participants. The mean difference was −0.82 kg, with a 95% confidence interval from −1.92 to 0.28 kg. The certainty of this evidence was rated moderate.[9] The minus sign favors the low GI or GL diets. The interval, however, stretches from a larger decrease in that group to a small advantage for the comparison group.

For this weight comparison, the interval includes zero, leaving the direction of the difference uncertain.[9] Reporting only −0.82 kg would remove the part of the result that explains why the authors were cautious. Both the estimate and its interval belong in the answer to “does this diet help more?” The first gives the average difference observed; the second helps show how firmly that difference can be interpreted.

Against other types of diet, the mean weight difference was −1.24 kg, with a 95% confidence interval from −2.82 to 0.34 kg. That analysis included 3 studies and 723 participants. Again, the authors concluded that there was probably little to no difference in weight change.[9] Changing the comparison category did not turn the evidence into a clear promise of greater weight loss.

These are differences between groups, rather than the total amount lost by everyone following a low GI diet. A small difference between diets does not mean that no participant lost weight. People in both groups may lose weight while the difference between groups remains uncertain.[9] If your own weight decreased after changing your meals, this review does not erase that experience. It asks whether the low GI condition provides a reliably greater benefit than the alternative.

A separate 2022 meta-analysis combined 10 studies in adults with excess weight. Across adults with a body mass index, or BMI, of 25 or above, low GI or GL diets were not shown to be superior for weight reduction. In the analysis restricted to adults with BMI of 30 or above, the mean difference was −0.93 kg, with a 95% confidence interval from −1.73 to −0.12 kg.[1] The BMI thresholds distinguish the broader excess-weight population from the obesity-only analysis; BMI relates weight to height squared.[1][8]

The restricted result deserves attention without replacing the overall result. The review considered adults with BMI of 25 or above and adults with BMI of 30 or above separately.[1] It would be misleading to quote the favorable subgroup as though it described every adult seeking weight loss. The population named next to the number is part of the finding, rather than a detail to discard when shortening the headline.

That meta-analysis found no differences between dietary groups in fat mass, fat-free mass, or waist circumference.[1] Fat-free mass means the amount of body tissue and other components apart from fat. The review thus contains a weight advantage in a restricted analysis alongside a lack of differences in measures of body composition. A result about weight cannot automatically answer what happened to body fat or the rest of the body.

Why older and newer reviews reach different answers

The 2007 review offered a more favorable weight loss conclusion. It included 6 randomized controlled trials involving 202 participants. The pooled weight analysis covered 163 people and found a mean difference of −1.1 kg in favor of low GI or GL diets, with a 95% confidence interval from −2.0 to −0.2 kg. Interventions lasted from 5 weeks to 6 months, and the authors called for longer follow-up.[20]

There was a reason for the early interest: those trials found greater weight loss with the lower GI or GL diets. But their duration also limited the question they could answer.[20] An encouraging result over the periods studied is not an answer about how a diet will perform indefinitely. Reading the positive conclusion together with the request for longer follow-up gives a more accurate picture of what the review established at the time.

The 2023 update was not simply the same collection with more participants added. Its eligibility criteria changed. Only 1 study from the previous version was retained, while 9 studies were newly identified. The minimum intervention duration was set at 8 weeks.[9] The collection of evidence changed as well as its date. Different study sets can lead to different conclusions; the update asked the question under revised conditions.

The larger 2018 review included 101 studies and 8,527 participants. It reported small overall improvements in weight and BMI. Yet when results were examined by the type of comparison diet, no individual control diet differed significantly from low GI diets.[15] Combining a broad range of comparisons can produce a different answer from asking whether low GI eating outperforms one particular alternative.

There was also a favorable finding in studies of people with normal blood glucose that achieved a dietary GI difference of at least 20 points. The standardized mean difference in weight was −0.26, with a 95% confidence interval from −0.43 to −0.09.[15] A standardized mean difference puts study results onto a common scale. It does not mean a difference of −0.26 kg.

This finding also draws attention to the diet participants actually managed to follow. A plan described as low GI and a substantial achieved difference in dietary GI are not interchangeable descriptions. The review emphasized the need for strategies that help people follow the diet in practice.[15] The practical challenge belongs inside the interpretation of the results: naming the intended diet does not fully describe the dietary contrast that was achieved.

A useful reading order is to look at the overall weight comparison first, then examine the narrower populations and conditions. The updated review rated evidence across its outcomes from moderate to very low certainty.[9] Starting only with a favorable subgroup can hide the uncertainty in the broader picture. Counting positive and negative papers as votes is less useful than asking who participated, what diets were compared, and how much those diets differed.

Numbers: low GI diets in weight loss trials

−0.82 kg: the average difference in weight change against higher GI or GL diets; the 95% confidence interval was −1.92 to 0.28 kg and included no difference.[9]

−0.93 kg: the weight difference in the analysis restricted to adults with BMI of 30 or above; the 95% confidence interval was −1.73 to −0.12 kg.[1]

−1.1 kg: the weight difference in the 2007 review’s analysis of 163 participants; the 95% confidence interval was −2.0 to −0.2 kg.[20]

12 weeks: the duration of a comparison of dietary patterns in 129 young adults with overweight or obesity; differences in percentage weight loss between groups were not statistically significant.[8]

−0.31 percentage points: the difference in HbA1c in the diabetes analysis, with a 95% confidence interval from −0.42 to −0.19; this is a blood sugar management result, not an amount of weight lost.[2]