Research from the Nurses’ Health Study II (NHSII) suggests an association between pre-pregnancy potato consumption and subsequent gestational diabetes mellitus (GDM).

The NHSII recruited a cohort of 116,430 nurses, aged 24 to 44, in 1989 in an ongoing study. Between 1991 and 2001, 21,693 singleton pregnancies were identified in 15,632 women, of which 854 (5.5%) were complicated by incident GDM.

Using food frequency questionnaires completed every two years it was possible to identify potato products consumed and quantify their frequency of consumption. These data were compared with a comprehensive set of demographic, nutritional and lifestyle variables. Robust statistical analyses including testing with multiple regression analyses models was performed.

Four levels of potato consumption were defined: < 1 serving per week;

1 serving per week; 2-4 servings per week and ≥ 5 servings per week.

The statistical model controlled for age, parity, ethnicity, family history of diabetes, smoking, energy intake and physical activity. When compared with < 1 serving per week, the relative risks (RR) of developing GDM were:

• RR 1.25 (95% CI: 1.01-1.54)

for 1 serving/week

• RR 1.45 (95% CI: 1.19-1.77)

for 2-4 servings/week

• RR 1.81 (95% CI: 1.39-2.37)

for ≥ 5 servings/week; P < 0.001 for trend.

A further analysis was performed to estimate the effect of substituting a non-starchy food for two servings of potato per week. The reduction in the risk of GDM had the following substitutions been made was:

• 9%, RR 0.91 (95% CI: 0.85-0.97)

for other vegetables

• 10%, RR 0.90 (95% CI: 0.83-0.99)

for legumes

• 12%, RR 0.88 (95% CI: 0.83-0.94)

for whole grain foods

Along with pasta, bread and rice, potato is a starchy food with a high glycaemic index (GI). High GI diets are associated with raised fasting glucose levels, insulin resistance and type 2 diabetes.

Observational studies always need to be interpreted with great caution. As the authors concede, association does not prove causation. This is particularly so when, as in this paper, statistical estimates of how different eating patterns might have modified those associations are presented as if they were actual interventions.

Having said that the NHSII is a leading nutritional epidemiology programme and the findings presented here are eminently plausible; that highly glycaemic food consumption is directly linked to metabolic disorders such as GDM.

Dr Chris Barclay