Eating white rice may be linked to a higher risk of Type 2 diabetes.
A review of four studies involving around 350,000 people found the more white rice people ate, the higher their chance of developing the condition seemed to be.
Experts from Harvard University in Massachusetts looked at two studies in Asian people and two in Western populations.
Asian people tended to have a much higher white rice intake than those in the West, averaging three to four servings a day compared with one or two servings per week.
The results showed that Asian people have a higher chance of developing Type 2 diabetes, with those who ate the most at highest risk.
"What we've found is white rice is likely to increase the risk of Type 2 diabetes, especially at high consumption levels such as in Asian populations," Qi Sun of the Harvard School of Public Health told AFP.
"But at the same time people should pay close attention to the other things they eat.
"It's very important to address not just a single food but the whole pattern of consumption."
In the British Medical Journal (BMJ), Mr Sun's team said the link emerged from an analysis of four previously published studies, carried out in China, Japan, Australia and the United States.
These studies followed 350,000 people over a timescale from four to 22 years. More than 13,000 people developed Type 2 diabetes.
In the studies carried out in China and Japan, those who ate most rice were 55% likelier to develop the disease than those who ate least. In the US and Australia, where consumption of rice is far lower, the difference was 12%.
Participants in the two Asian countries ate three or four servings of rice a day on average, compared to just one or two servings a week in the Western countries.
White rice is the dominant form of rice eaten in the world. Machines produce its polished look by hulling and milling, leaving a grain that is predominantly starch.
Brown rice, by contrast, has more fibre, magnesium and vitamins, and a lower "glycaemic index," a measurement of sugar content, than white rice.
Mr Sun said the study did have limitations, including full details about what the volunteers ate in addition to rice.
"I don't think I can put forward a 100% confirmed case, given that this is a meta-analysis of four original studies," he said.
"But I see a consistency across these studies, and there is biological plausibility that supports the association between white rice consumption and diabetes.
He added: "More trial data are needed to corroborate or refute our observations."
Diabetes affects nearly 350 million adults worldwide, according to the US Centers for Disease Control and Prevention (CDC).
