• % of patients who adhere to the diet as recorded in food diaries [ Time Frame: 52 weeks ]
  Adherence to the diet will be assessed by self report food diaries which patients will keep. These will be assessed at 12 weeks and 52 weeks after beginning the diet.
• Score on diet adherence likert scale questionnaire [ Time Frame: 52 weeks ]
  Adherence to the diet will be assessed by patients scores on a self report likert scale, ranging from 1 (follow the diet all the time) to 5 (follow the diet none of the time), at 12 weeks and 52 weeks after beginning the diet.
• Score on diet acceptability likert scale questionnaire [ Time Frame: 52 weeks ]
  Acceptability of the diet will be assessed by patients scores on a self report likert scale, ranging from 1 (the diet is acceptable all of the time) to 5 (the diet is acceptable none of the time), at 12 weeks and 52 weeks after beginning the diet.

• Proportion of patients showing a 50% or greater reduction in seizure frequency from baseline [ Time Frame: 52 weeks ]
  Seizure frequency will be assessed through patient completion of seizure diaries throughout the study. This will be assessed at baseline and 12 weeks, 24 weeks and 52 weeks after beginning the diet.
• Change from baseline in QOLIE-31 score [ Time Frame: 52 weeks ]
  Epilepsy related quality of life will be measured using the Quality of Life in Epilepsy Inventory-31 (QOLIE-31), a self report questionnaire comprising two factors (emotional and psychological effects, and medical and social effects), eight sub-scales, and 39 items. Items are measured on 4- to 6- point Likert scales, with a maximum total score of 100. Higher scores indicate a better QoL. Change in QOLIE-31 score from baseline to 12 week and 52 weeks after beginning the diet will be assessed.

Many persons with epilepsy have seizures which remain uncontrolled by anti-epilepsy medications and are unsuitable for or unwilling to undergo surgical treatments for their epilepsy, or have undergone such treatments and continue to have seizures. Dietary treatments for epilepsy have been shown to be effective in children, and are probably effective in adults, but compliance with the classic ketogenic diet (KD) and to some degree also the modified Atkins diet (MAD) seems difficult for many adults.

The LGI diet may be easier and in children appears to be of comparable efficacy to other dietary treatments (KD and MAD), but has been little studied in adults.

This is a randomised study of immediate versus deferred LGI diet in adults with seizures incompletely controlled by anti-epilepsy medications. 12 weeks of dietary treatment in those randomised to LGI will be followed by the opportunity for the control group to undertake 12 weeks of the LGI diet.

• Experimental: Immediate Diet Group
  Adhere to Low Glycaemic Index diet for 52 weeks, beginning immediately after randomisation.
  Intervention: Other: Low Glycaemic Index Diet
• Active Comparator: Delayed Diet Group
  Adhere to Low Glycaemic Index diet for 52 weeks, beginning 12 weeks after randomisation.
  Intervention: Other: Low Glycaemic Index Diet

Inclusion Criteria:

• adults (>18 years)
• epilepsy refractory to at least two appropriate first line anti-epileptic drugs in adequate dose, unsuitable for or unwilling to undergo surgery for epilepsy.
• At least 1 seizure per week.
• BMI >18

Exclusion Criteria:

• unable to give informed consent
• unable to comply with diet
• surgery for epilepsy or VNS within the last 12 months
• non-epileptic seizures
• pregnant or planning pregnancy
• significant renal impairment
• history of renal stones