White rice consumption and risk of type 2 diabetes: meta-analysis and systematic review

BMJ 2012; 344 doi: http://dx.doi.org/10.1136/bmj.e1454 (Published 15 March 2012)
Cite this as: BMJ 2012;344:e1454

Recent rapid responses

Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on bmj.com. Although a selection of rapid responses will be included as edited readers' letters in the weekly print issue of the BMJ, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window.

Displaying 1-10 out of 18 published

Neeta Kumar1, Neeru Gupta1, Pratik Kumar2, Jugal Kishore3
1. Indian Council of Medical Research, India
2. All India Institute of Medical Sciences
3. Maulana Azad Medical College, India

Using the word “Asian” while quoting the effect of white rice from analyzing studies done in metro cities of Japan and Shanghai is not adequate. The authors have failed to consider the very fact that the quality of rice varies in varying regions of Asia, which is not merely a city in Japan or China. There are hundreds of strains in the whole of Asia of white rice as well as brown rice.

The glycemic index of a food item is the measure of its physiologic impact on the body to produce insulin. If GI is more the food is considered high glycemic and risky for producing Diabetes. The Glycemic index of brown rice has been found to be more than that of white rice on page no 18-21 of the article in reference number one.

“Many people have raised concerns about the variation in published GI (glycemic Index) values for apparently similar foods. This variation may reflect both methodologic factors and true differences in the physical and chemical characteristics of the foods. One possibility is that 2 similar foods may have different ingredients or may have been processed with a different method, resulting in significant differences in the rate of carbohydrate digestion and hence the GI value. Two different brands of the same type of food, such as a plain cookie, may look and taste almost the same, but differences in the type of flour used, in the moisture content, and in the cooking time can result in differences in the degree of starch gelatinization and consequently the GI values. International table of glycemic index. Rice, for example, shows a large range of GI values, but this variation is due to inherent botanical differences in rice from country to country rather than to methodologic differences. Differences in the amylose content could explain much of the variation in the GI values of rice (and other foods) because amylose is digested more slowly than is amylopectin starch. GI values for rice cannot be reliably predicted on the basis of the size of the grain (short or long grain) or the type of cooking method. Rice is obviously one type of food that needs to be tested brand by brand locally. Carrots are another example of a food with a wide variation in published GI values.

Indian studies do not imply particular food item for increased Diabetes risk (2) since a variety of food, variety of cooking methods, socioeconomic conditions, quality of life, sun exposure, mental stress, physical activity are applied in the region. So Indian scientists find 'Fast food culture’ and ‘Sedentarinism’- the main drivers of diabetes epidemic in India (3).

1. International table of glycemic index and glycemic load values: 2002,Kaye Foster-Powell, Susanna HA Holt, and Janette C Brand-Miller. Am J Clin Nutr January 2002 vol. 76 no. 1 5-56

2. Risk factors and complications of type 2diabetes in Asians. Rajbharan Yadav, Pramil Tiwari* and Ethiraj Dhanaraj. Review Article CRIPS Vol. 9 No. 2 April-June 2008.

3. Epidemiology of type 2 diabetes: Indian scenario. V. Mohan, S. Sandeep, R. Deepa, B. Shah. Indian J Med Res 125, March 2007, pp 217-230

Competing interests: None declared

Neeta Kumar, Medical Scientist

Neeru Gupta, Pratik kumar, Jugal Kishore

Scientist, Indian Council of Medical Research, Ansari Nagra, New Delhi-110029

Click to like:

Dear Editor:
We read with interest Oster et al’s response and the table. However, we were puzzled by the data presented in this poorly-denoted table. We did not understand why adjusted means were not provided for all cells. Further, the 95% CIs did not necessarily include the means listed in the table. Nonetheless, in response to Oster et al’s concerns on the multivariate adjustments in the studies included in our meta-analysis,1 we want to emphasize that age and body mass index (BMI) were controlled for in all studies. In addition, a wide array of other variables was also adjusted in these studies. Besides the multivariate adjustments, the validity of the data was also underpinned by the prospective study design, use of validated questionnaires to assess diet, and high follow-up rate. On the contrary, we were concerned about the validity of Oster et al’s data. Given the positive correlations of age/BMI with rice intake and glucose levels, we shall anticipate that the inverse association between rice intake and fasting glucose would be strengthened after adjusting for age and BMI, but the opposite was shown in the table. In addition, Oster et al described that the correlation between BMI and fasting glucose was strong, which should be expected. However, the adjusted means were almost identical between overweight and obese participants.

Some key questions are yet to be addressed in this research. For example, more data are warranted to examine the effects of replacing white rice with brown rice on insulin resistance, and it is interesting to further explore the biological mechanisms underlying the association between white rice and diabetes risk.

1. Hu EA, Pan A, Malik V, Sun Q. White rice consumption and risk of type 2 diabetes: meta-analysis and systematic review. BMJ 2012;344:e1454.

Competing interests: None declared

Qi Sun, Nutritional Epidemiologist

Harvard School of Public Health; Brigham and Women's Hospital and Harvard Medical School, 665 Huntington Avenue, Boston, MA 02115, USA

Click to like:

Dear Editor:
In our meta-analysis on white rice consumption and diabetes risk,1 we included more than 13,000 newly-diagnosed type 2 diabetes patients identified from more than 350,000 participants in four studies. In epidemiological studies of such a large scale, monitoring all participants for blood glucose levels is simply infeasible. In addition, we want to further stress that the effects of inclusion of undiagnosed diabetes during follow-up are likely to be small and tend to bias the results toward the null. The self-reported diabetes diagnosis has been demonstrated to be accurate in the US and Japanese studies.2 3 In the Chinese study, high rice consumption was correlated with lower socioeconomic status,4 which may lead to under-diagnosis of diabetes, although such a bias is most likely to attenuate true associations.

We followed a standard and comprehensive protocol in searching relevant literature, and we did not apply any restriction on language. Identification of unpublished data is infeasible, if not impossible. The quality of such data is difficult to examine in that those data are not evaluated by peer-review.

Data quality assessment is critical to meta-analyses. We followed current guidelines and customized quality assessment criteria by putting much weight on the assessment of diet, which is fundamental to all nutritional epidemiological research.

The significant heterogeneity that Naqvi et al noted was due to ethnicity. We, therefore, reported pooled estimates for Asian and Western populations, respectively.

Lastly, Naqvi et al erroneously used cumulative incidence or risk to describe prevalence. Apparently, the participants included in these studies were not representative of general population in each country, let alone the whole world. Although this unrepresentativeness may limit generalizability, it does not necessarily affect the internal validity of the results.

1. Hu EA, Pan A, Malik V, Sun Q. White rice consumption and risk of type 2 diabetes: meta-analysis and systematic review. BMJ 2012;344:e1454.
2. Manson JE, Colditz GA, Stampfer MJ, Willett WC, Krolewski AS, Rosner B, et al. A prospective study of maturity-onset diabetes mellitus and risk of coronary heart disease and stroke in women. Arch Intern Med 1991;151(6):1141-7.
3. Nanri A, Mizoue T, Noda M, Takahashi Y, Kato M, Inoue M, et al. Rice intake and type 2 diabetes in Japanese men and women: the Japan Public Health Center-based Prospective Study. Am J Clin Nutr 2010;92(6):1468-77.
4. Villegas R, Liu S, Gao YT, Yang G, Li H, Zheng W, et al. Prospective study of dietary carbohydrates, glycemic index, glycemic load, and incidence of type 2 diabetes mellitus in middle-aged Chinese women. Arch Intern Med 2007;167(21):2310-6.

Competing interests: None declared

Qi Sun, Nutritional Epidemiologist

Harvard School of Public Health; Brigham and Women's Hospital and Harvard Medical School, 665 Huntington Avenue, Boston, MA 02115, USA

Click to like:

Hu's and colleagues metaanalysis demonstrates an association between rice intake and diabetes risk. The authors note that "most established risk factors for type 2 diabetes were adjusted for in the fully adjusted models in these studies." We recently completed a study of 621 Chinese immigrants living in New York City who attended free cardiovascular screening, daily number of times rice was inversely, though not statistically correlated with fasting glucose level. Daily rice servings were correlated with age and BMI, both of which were strongly correlated with fasting glucose(See table). While our findings are limited by their cross-sectional design and much smaller sample size, it is impossible to judge the validity of the Hu study without more information regarding how adjustments for a likely colinearity between rice servings, age, and BMI.

Competing interests: None declared

Ady Oster, Physician

Charles B. Wang Community Health Center, 268 Canal Street New York, NY 10013 USA

Click to like:

Emily Hu’s recent paper further endorses the significant role that the consumption of white rice plays in the development of type 2 diabetes. Whilst acknowledging the importance of this publication, it also brings into focus the urgent need to develop practical strategies to reduce the glycemic load of rice based diets. In most Asian countries, rice remains the major staple and is unlikely to be displaced by any other staples. Per capita rice consumption in Asia ranges between 70- 245 kg/year and in China and India consumption rates are 77 and 71 kg/year respectively (FAOSTAT, 2012). In societies with a long tradition of rice consumption, dietary advocacy to reduce its intake will remain an untenable goal. Strategies to reduce the glycemic load of rice based diets should therefore become a priority. Since rice is rarely eaten alone but in combination with animal or plant based foods, the challenge is to develop dishes and condiments to accompany rice that will reduce the glycemic load. Other strategies include the use of plant breeding and processing techniques such as parboiling. Translating these epidemiological observations into practical solutions will be the greatest public health challenge ahead.


FAOSTAT, 2012. Per Capita Rice consumption FAO, Rome.

Competing interests: None declared

C. Jeya K. Henry, Professor (Human Nutrition)

Viren Ranawana

Singapore Institute for Clinical Sciences, 30, Medical Drive, Singapore 117609

Click to like:

Substitution of whole grains, including brown rice, for white rice may lower the risk of type 2 diabetes.[1] This has environmental impacts as well, as rice is the staple food for nearly two-thirds of the world's population. A change in consumption patterns from per boiled rice to untreated rice may also reduce the risk of arsenic contamination in arsenic prone areas. [2] It has been tried and the main barriers to acceptance were noted to be the perception of rough texture, unpalatable taste and higher price of brown rice. [3] Perhaps increased awareness and some incentives may be required for achieving a change of widespread dietary habit of eating white rice.[1]

[1] Sun Q, Spiegelman D, van Dam RM, Holmes MD, Malik VS, Willett WC, Hu FB. White rice, brown rice, and risk of type 2 diabetes in US men and women. Arch Intern Med. 2010;170(11):961-9.
[2] Roy P, Orikasa T, Okadome H, Nakamura N, Shiina T. Processing conditions, rice properties, health and environment. Int J Environ Res Public Health. 2011;8(6):1957-76.
[3] Zhang G, Malik VS, Pan A, Kumar S, Holmes MD, Spiegelman D, Lin X, Hu FB. Substituting brown rice for white rice to lower diabetes risk: a focus-group study in Chinese adults. Am Diet Assoc. 2010;110(8):1216-21.

Competing interests: None declared

Debasish Debnath, Specialist Registrar, Department of surgery

Frimley Park Hospital, Frimley

Click to like:

Many developed countries see a large increase in their prevalence of obesity and T2D.

Weight reduction programmes aim to reduce unhealthy body fat. Insulin stimulates fat production making weight loss difficult or impossible when insulin is the predominant metabolic hormone.

In pre T2 Diabetes conditions there is glucose intolerance which is often characterised by high circulating insulin levels and insulin resistance. A similar picture exists in the early stages of T2D but as B cell function deteriorates and if insulin resistance is not reduced the bodies coping mechanisms are overstretched and blood glucose levels rise.

Insulin resistance is difficult to measure and methods such as the euglycaemic clamp simply serve to measure resistance to one of it's many functions. Perhaps functions such as stimulating fat production are less resistant and increased fat production ensues.

For these reasons weight loss in T2 diabetes should be facilitated by lowering the bodies glucose led drive to make Insulin. This is largely fuelled by a carbohydrate rich diet and a lack of exercise to use up such energy.

Carbohydrate counting regimes, low GI diets and simple restriction of carbohydrates are sensible ways of reducing the glycaemic load on the body and controlling Insulin production and glycaemic control without the need for continuingly complex regimes of hypoglycaemic agents. Metformin, glitazones, exercise and weight loss are already known to reduce progression of IGT to diabetes - and all will lower endogenous insulin production. A recent study also suggested that diabetes can be cured by significant calorie restriction and weight loss.

The specific use of lowering of dietary carbohydrate to help to control diabetes is seldom advocated to patients by dietitians and clinicians alike. Yet there is sound physiological reasons why doing so would lower insulin production and reduce the harmful effects that too much circulating insulin can cause.

Biochemists and endocrinologists agree with this physiology yet no one seems to put it into practical use in a clinical setting. It is time that there was a sensible discussion and further research on this important topic. As Gary Taubes eloquently puts it in his well researched novel , "the diet delusion" - the low fat diet remains the ingrained panacea for weight loss.

It is a leap of faith but could a carbohydrate rich diet - in association with a susceptible genetic background - actually "cause" T2D? Certainly epidemiology would suggest that in some populations the prevalence of T2D has risen dramatically as the industrial revolution and similar non physiological forces have made the consumption of carbohydrate (and fat) by the population increase. Food for thought?

Competing interests: None declared

Ian Dickson, GP

Springwell Medical Centre, 39 Ardmillan Terrace, Edinburgh EH11 2JL

Click to like:

Rapid Response[i] :

The author’s main objective was to assess the association between white rice consumption and risk of type 2 diabetes. The primary outcome analysed in the review was “risk of type 2 diabetes”. However, the authors did not clearly report the methods used to measure the primary outcome which was stated as self reported. In table 2 there is some indication of type 2 diabetes confirmation by medical practitioners. On closer inspection this confirmation by medical testing was only carried out on those subjects who reported positive for diabetes by self report and not all subjects.  


A comprehensive search strategy was not reported in the review, hence, it was not possible to assess if all relevant studies were included. The decision to include only published prospective cohort studies in English may have introduced bias into this systematic review via the exclusion of unpublished studies and those in other languages. The Data Quality Assessment appeared to have been devised by the authors of the review and was not a standard tool for critical appraisal.[ii] [iii]


The authors found that three studies showed a statistically significant association between high intake of white rice and risk of type 2 diabetes.  The authors reported the results of included studies to be significantly heterogeneous (Overall: I2  72.2%, P= 0.001). However, they did not discuss the possible reasons of heterogeneity in their analysis and proceeded to pool the findings for analysis.


In part the heterogeneity of the results might be explained by the range of prevalence reported in Table 1.  Five of the seven studies reported prevalence up to 2.5%. Studies from Japan and Australia reported prevalence of 1.4% and 1.15% respectively. Given that global prevalence for diabetes worldwide was estimated by the WHO as 2.8% in the year 2000 and is predicted to rise to 4.4% by 2030,[iv]  the reported low prevalence in these studies could be due to the fact that diabetes was self reported in these studies and thus did not reflect the actual prevalence of the disease. What is clear from this review is that future studies with robust methodologies will be essential to investigate any association between white rice consumption and risk for type 2 diabetes.



Table 1 (Taken from table 1 in the systematic review)






Hodge et al 2004 (41.1% male)





Villegas et al, 2007 (Female only)





Sun et al, 2010 (Male only)

United States




Sun et al, 2010 (Female only)

United States




Sun et al, 2010 (Female only)

United States




Nanri et al, 2010 (Male only)





Nanri et al, 2010 (Female only)








It was a group work done in the workshop “How to Practice Evidence-Based Health Care”, Oxford on 28th March 2012. By: Syed Ali Raza Naqvi*, Sarah Westwater-Wood*, Ghada  Alarfaj**, Helen Atherton**, Charles Cachoeira**, Joseph Marcel El Khoury**, Mary Ann McLane**, Laurence Pollissard-Badroy**, Liubov Yashina**, Carl Heneghan***.



* Co-facilitators

** Participants

*** Facilitator




[i] In the critical appraisal workshop (How to Practice Evidence-Based Health Care), at Oxford, we critically appraised the systematic review by Hu and colleagues as part of a group exercise.  (Centre for Evidence Based Medicine, How to Practice EBHC (Evidence-Based Health Care) [online] http://www.cebm.net/index.aspx?o=6562)


[ii] Public Health Resources U. Critical Appraisal Skills Programme (CASP), 2007.


[iii] Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions, 2011


[iv] Wild et al ., 2004, Global prevalence of diabetes. DIABETES CARE,(27) 5 , 1047-1053 [online] http://www.who.int/diabetes/facts/en/diabcare0504.pdf

Competing interests: None declared

Syed Ali Raza Naqvi, Research Officer

Sarah Westwater-Wood, Ghada Alarfaj, Helen Atherton, Charles Cachoeira, Joseph Marcel El Khoury, Mary Ann McLane, Laurence Pollissard-Badroy, Liubov Yashina, Carl Heneghan

Evidence Adoption Centre, Douglas House, 18 Trumpington Road, Cambridge, CB2 8AH.

Click to like:

The type of rice affects its Gycemic Index (GI), and it is high GI rice that is the problem. Basmati rice is only medium GI, so a moderate intake of basmati is safe.

See this:

It is a shame that this sort of research paper can get into BMJ and make headline news. It is like saying that some road accidents are caused by blue cars, and therefore if we avoid driving blue cars the roads will be safer.

Please ask the authors to take a "Statistics and interpreting causality 101" course.


Competing interests: None declared

Brian Wernham, Analyst

None, London, UK

Click to like:

The magnitude of this study, analyzing cases in four countries and encompassing 352,000 people is impressive. But its methodology raises more questions than it answers. This is because the universal nutritional paradigm upon which the research is based is flawed. Humans digest with greatest efficiency, the foodstuffs most common to their ancestral diets, whether that be rice, wheat or whale blubber. As a nutritional historian, I have found that the rates of obesity and Type II diabetes in African-Americans skyrocketed after this population group shifted from an ancestral diet of rice, yams, vegetables and poultry to wheat gluten, potato starch, pork and dairy products. If the researchers in this white rice and diabetes II study review their data, what they might find is that the real culprit is not "white rice," but rather the Asian addition of more European foodstuffs like wheat gluten and dairy products into their traditional rice diet.

Competing interests: None declared

Constance Hilliard, Nutritional Historian


University of North Texas, 1155 Union Circle #310650

Click to like: