Table 1.
Table summarising the results of changes in HbA1c from the eleven dietary interventions included in the systematic review (Mean values and standard deviations for mean weight loss and mean reduction HbA1c)
| HbA1c (%) at baseline | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| References | Participants | Mean | sd | Intervention (n per arm) | Composition of prescribed diets | Mean weight loss (kg) | Mean decrease in HbA1c (%) | Duration | Attrition rate | Medication | Conclusion |
| Guldbrand et al.( 18 ) | 61 overweight and obese adults with type 2 diabetes | 7·35 | LFD (31) v. LCD (30) | LFD: 55–60 % CHO, 10–15 % protein, 30 % fat LCD: 20 % CHO, 30 % protein, 50 % fat | 2·97 v. 2·34 (P=0·33) | −0·2 v. 0: no significant difference between groups (P=0·76) | 2 years | 9·7 % in LFD group and 13·3 % in LCD | Authors reported that at 6 months there was a statistically significant difference in mean insulin dose in favour of the LCD (P=0·046) | No significant difference between dietary interventions in improving HbA1c | |
| Brehm et al. ( 19 ) | 124 overweight and obese adults with type 2 diabetes | 7·3 | High-MUFA diet (43) v. HC diet (52) | MUFA: 45 % CHO, 15 % protein, 40 % fat (20 % MUFA) CHO: 60 % CHO, 15 % protein, 25 % fat | 4·0 (sd 0·8) v. 3·8 (sd 0·6) (P=0·867) | 0 v. –0·1: authors stated no significant difference between groups (no P value reported) | 12 months | 31 % for the MUFA group and 16 % for the HC group | Authors reported a lack of available information about participant’s drug usage. Only information on 32 participants’ drug use was available, which showed no systematic differences between diet groups. Therefore no adjustments were made for glucose-lowering medication | No significant difference between dietary interventions in improving HbA1c | |
| Fabricatore et al.( 9 ) | 79 obese adults with type 2 diabetes | 6·8 | Low fat (39) v. low GL (40) | Low fat: <30 % fat Low GL: 3 or less servings of moderate-GL and 1 or less serving of high-GL foods/d | 4·5 (sd 0·34) v. 6·4 (sd 0·52) (P=0·28) | 0·1 (sd 0·0012) v. 0·8 (sd 0·0104): significant difference between groups (P=0·01) | 40 weeks | 36·7 % | Authors stated that changes in HbA1c were adjusted for medication use. Percentage of participants who increased, decreased or did not change their diabetic medication regime did not differ between the groups at week 20 (P=0·51) or at week 40 (P=0·70) | Low GL appears to be more effective in reducing HbA1c compared with a the LFD | |
| Elhayany et al.( 12 ) | 259 overweight and obese adults with type 2 diabetes | 8·3 | LCMD (61) v. TM diet (63) v. ADA diet (55) | LCM: 35 % low-GI CHO, 15–20 % protein, 45 % fat rich in MUFA TM: 50–55 % low-GI CHO, 15–20 % protein, 30 % fat rich in MUFA ADA: 50–55 % CHO, 15–20 % protein, 30 % fat | 10·1 v. 7·4 v. 7·7 authors stated no significant difference between groups (no P value reported) | 2·0 v. 1·8 v. 1·6 significant difference between diets (P=0·021), LCM different than ADA, TMD different than ADA | 12 months | 30·9 % | Authors do not mention baseline medication characteristics or any changes in glucose-lowering medication use during the course of the intervention | LCM diet appears to be more effective in reducing HbA1c compared with a TM and ADA diets | |
| Barnard et al.( 13 ) | 99 obese adults with type 2 diabetes | 7·99 | Low-fat vegan diet (49) v. ADA diet (50) | Low-fat vegan diet: 75 % CHO, 15 % protein, 10 % fat ADA: 60–70 % CHO, 15–20 % protein and MUFA | 4·4 (sd 0·9) v. 3·0 (sd 0·8) (P=0·25) | 0·34(sd 0·19) v. 0·14 (sd 0·17) no significant difference between groups (P=0·43) 0·4 % v. 0·01 % (P=0·03) when adjusted for medication | 74 weeks | 18·4 % for the vegan group 14 % for the ADA group | Net 74-week dosages were reduced in 35 % participants in vegan group and 20 % of those in the ADA group, and were increased in 14 % of vegan group and 24 % of conventional group | Once data are adjusted for medication use, there appears to be a significant benefit in the low-fat vegan diet in decreasing HbA1c compared with the ADA diet | |
| Esposito et al.( 10 ) | 215 overweight adults with type 2 diabetes | 7·73 | LCMD (107) v. LFD (108) | LCMD: 50 % CHO, 20 % protein, no <30 % fat LFD: no >30 % fat with no >10 % SFA | 3·8 (sd 2) v. 3·2 (sd 1·9) authors stated no significant difference between groups (no P value reported) | 0·9 (sd 0·6) v. 0·5 (sd 0·4) authors stated significant differences between groups (no P value reported) | 4 years | 9·3 % | After 4 years 44 % of participants in the LCMD and 70 % of those in the LFD group required treatment (absolute difference, −26·0 percentage points (95 % CI 0·51, 0·86), hazard ratio adjusted for weight change, 0·70 (95 % CI 0·59, 0·90); P<0·001). | LCMD appears to be more effective in reducing HbA1c compared with a LFD with less need for glucose-lowering medication | |
| Pedersen et al.( 14 ) | 76 overweight adults with type 2 diabetes | 7·3 | HPD (21) v. standard protein diet (24) | HPD: 40 % CHO, 30 % protein, 30 % fat SPD: 50 % CHO, 20 % protein, 30 % fat | 9·7 (sd 2·9) v. 6·6 (sd 1·4) (P=0·32) | 0·4 v. 0·3 no significant difference between groups (P=0·29) | 12 months | 40·8% | Did not account for changes in medication, although authors stated that 4 volunteers managed their diabetes with diet alone, and all others treated with oral medication and/or insulin | No significant difference between dietary interventions in improving HbA1c | |
| Iqbal et al.( 17 ) | 144 obese adults with type 2 diabetes | 7·75 | LCD (40) v. LFD (28) | Low CHO: <30 g/d CHO Low fat: ≤30 % of energy for fat with <7 % of energy from SFA | 1·5 v. 0·2 (P=0·147) | 0·1 v. 0·2 no significant difference between groups (no P value reported) | 24 months | 60 % in the low-CHO group and 46 % in the low-fat group | Authors stated that many participants were unable to provide information regarding changes to medication or dosages and therefore the effects of glucose-lowering medication was not adjusted for | No significant difference between dietary interventions in improving HbA1c | |
| Ma et al.( 11 ) | 40 overweight and obese adults with type 2 diabetes | 8·42 | ADA diet (21) v. low-GI diet (19) | ADA: 60–70 % CHO, 15–20 % protein and 30 % fats Low GI: participants given goals to reduce daily dietary GI score to 55 | 0·80 v. 1·32 (P=0·89) | 0·43 v. 0·35 no significant difference between groups (P=0·88) | 12 months | 10 % for the ADA group and 10 % for the low-GI group | Participants in the low-GI group were less likely to add or increase dosage of glucose-lowering medications (OR 0·26; P=0·01) | No significant difference between dietary interventions in improving HbA1c | |
| Milne et al.( 15 ) | 70 overweight and obese adults with type 2 diabetes | 10·0 | 0·35 | Weight management diet (21) v. HC/fibre diet (21) v. modified lipid diet (22) | Weight management diet: restrict extrinsic simple sugars and energy dense foods, no advice for macronutrient contribution HC/fibre: 55 % CHO, 15 % protein, 30 % fat, 30 g or more dietary fibre/d Modified lipid diet: 45 % CHO, 19 % protein, 36 % fat | −1·5 v. 1·0 v. 0·1 authors stated no significant difference between groups (no P value reported) | 0·1 v. 0·1 v. 0·2 authors stated no significant difference between groups (no P value reported) | 18 months | 8·6 % | Authors state that 52·3 % of the weight management group, 57·1 % of the HC/fibre group and 50 % of the modified lipid diet were on glucose-lowering medication at baseline. However there is no mention of medication adjustments being made throughout the study | No significant difference between dietary interventions in improving HbA1c |
| Larsen et al.( 16 ) | 99 overweight and obese adults with type 2 diabetes | 7·84 | HPD (53) v. HC diet (46) | HP: 40 % CHO, 30 % protein, 30 % fat HC: 55 % CHO, 15 % protein, 30 % fat | 2·23 v. 2·17 (P=0·78) | 0·23 v. 0·28 no significant difference between groups (P=0·44) | 12 months | 9·2 % HP group v. 4·2 % HC | Authors reported a significant reduction in the requirement for glucose-lowering medication in the HP group compared with the HC group at 3 months (P=0·03) although the difference was no longer significant at 12 months (P=0·05). However authors stated that that there were not significant differences in the decrease in HbA1c between groups when values were adjusted for changes in medication | No significant difference between dietary interventions in improving HbA1c | |
LFD, low-fat diet; LCD, low carbohydrate diet; CHO, carbohydrate; HC, high carbohydrate; low GL, low glycaemic load; LCMD, low carbohydrate Mediterranean diet; TM, traditional Mediterranean; ADA, American Diabetes Association; LCM, low carbohydrate Mediterranean; GI, glycaemic index; TMD, traditional Mediterranean diet; HPD, high-protein diet; SPD, standard protein diet; HP, high protein.