Skip navigation

Evidence-based vegetables and diabetes


Bandolier is intrigued by high-quality studies associating behaviour, especially eating and drinking, with health outcomes. Such studies cannot prove a link, merely suggest an association. There may be confounding variables or factors which no-one has yet thought of, and links may not be causal.

But when consistent patterns emerge linking a type of behaviour with good health outcomes, and when the 'definitive' randomised trial may never be done, we have to ask ourselves when the evidence is sufficient for us to change our behaviour. That may include not only the strength of the evidence available, but also our own hopes, fears, and biases. We may fear cancer more than heart disease, or diabetes more than both.

There is bias in this. Bandolier is a fan of vegetables and salads (and chips, but don't mention that). So notice is taken of a study associating year-round consumption of salad vegetables with reduced risk of diabetes [1], especially when the study is large and good.


This was part of the Isle of Ely Study (run from Cambridge) which is prospectively studying the aetiology of type II diabetes. From a randomly-selected sample of the population age 40 to 64 years, and after excluding people with diabetes or who had moved away, 1122 people (73% of those asked to participate) were included in the study. They came to a screening centre and underwent a standard 75 g oral glucose tolerance test, had lots of questions asked and measurements made, and particularly answered a battery of questions about their food consumption.

Data on the frequency of consumption of foods was coded into the broad areas of frequent (daily consumption of food on most days) and infrequent. Fruit, salad and raw vegetables were divided into the broad groups of frequent all year, frequent in summer only, and infrequent all year.


There were 51 people (4.5%) who had non-insulin dependent diabetes mellitus (NIDDM) by WHO criteria, and 188 (17%) who had an impaired glucose tolerance test. Age-standardised rates were 2.3% and 11.2% respectively. These people were older, fatter and less active than those with normal glucose tolerance tests.

People who ate salad vegetables frequently all year had a lower incidence of impaired glucose tolerance (13%) than those who ate them less frequently (17%). Most impressive, though, was the incidence of NIDDM (Figure), which was only 1% in people eating salad vegetables frequently throughout the year, compared with almost 6% with those eating them infrequently throughout the year.

The results for salad vegetable consumption remained significant after various adjustments. It was apparent in people who were not overweight. Frequent fruit consumption was not significantly associated with lower diabetes incidence.


Another brick in the wall for evidence-based health eating and healthy living. Evidence associating vegetable consumption as protective against bad things happening to an individual's health continues to build. Eating vegetables seems to protect against heart disease and stroke, probably protects against some forms of cancer, and now is implicated in protecting against the development of diabetes.

It would be interesting to speculate whether significant shifts in healthy living behaviour would contribute to reduced healthcare demands. If cardiovascular disease, cancer and diabetes are major consumers of health care resources, and the incidence of these diseases fell because we all ate more vegetables, what would we die of? Would we just consume more social services instead? Someone must have modelled this.


  1. DE Williams et al. Frequent salad vegetable consumption is associated with a reduction in the risk of diabetes mellitus. Journal of Clinical Epidemiology 1999 52: 329-335.

previous or next story in this issue