Table 1.
Characteristics of included trials
| Trial | BMI; HbA1C range; insulin dependent inclusive? | Mean age (years: (intervention/control); % female | Carbohydrate intake; caloric restriction; intense behavioral support | Comparator (diet, iso-caloric) | Adherence | Missing participant outcome data | Duration of intervention (months) |
|---|---|---|---|---|---|---|---|
| Dyson et al (2010)34 | BMI >25; did not include patients receiving insulin (n=12) | 54 | <40 g/d; calorically restricted; intense behavioral support | Low fat | Adequate | 17% | 3 |
| Morris et al (2019)35 | BMI >30; did not include patients receiving insulin (n=33) | 69/64; 55% female | <26% carbohydrate/d; calorically restricted | Low fat | Adequate | 0% | 3 |
| Saslow et al (2014)36 * | BMI ≥25; did not include patients receiving insulin (n=34) | 65/55; 74% female | “A very low carbohydrate, high fat, non calorie-restricted diet whose goal is to induce nutritional ketosis”; intense behavioral support | Low fat; iso-caloric (per intake not per goal) | Adequate | 11% | 12 |
| Saslow et al (2017)37 | BMI≥25; did not include patients receiving insulin (n=25) | 53/58; 56% female | 20-50 g/d | Low fat; iso-caloric | Adequate | 24% | 8 |
| Yamada et al (2014)38 | Did include patients receiving insulin (n=24) | 63/63; 50% female | 70-130 g/d | Low fat; iso-caloric (per intake not per goal) | Adequate | 0% | 6 |
| Tay et al (2014)39 | BMI 26-45; HbA1c≥7.0% or taking diabetes medication; did include patients receiving insulin (n=131) | 58/58; 37% female | 50-70 g/d target; calorically restricted; intense behavioral support | Low glycemic index; iso-caloric | Adequate | 29% | 24 |
| Jönsson et al (2009)40 | HbA1c>5.5%; did not include patients receiving insulin (n=17) | 66/63; 24% female | <130 g/d | Low fat | Adequate | 18% | 6 (crossover study; first 3 month comparison used in analysis) |
| Sato et al (2017)41 | BMI >23; HbA1c>7.5%; did include patients receiving insulin (n=66) | 61/58; 23% female | 130 g/d target | Low fat | Adequate | 6% | 6 |
| Goldstein et al (2011)42 | BMI 30-40; HbA1C>7%; did not include patients receiving insulin (n=52) | 57/55; 48% female | 25-40 g/d; intense behavioral support | Low fat (ADA 2000); iso-caloric | Adherent at VLCD level | 42% | 12 |
| Guldbrand et al (2012)43 | Did include patients receiving insulin (n=61) | 61/62; 56% female | 20%; calorically restricted | Low fat; iso-caloric | Adequate | 0% | 24 |
| Iqbal et al (2010)44 | BMI≥30; did include patients receiving insulin (n=144) | 60/60; 11% female | <30 g/d; intense behavioral support | Low fat | Not adequate | 47% | 24 |
| Nishimori et al (2018)45 | NAFLD in addition; did include patients receiving insulin (n=28) | 49/50; 36% female | 70-130 g/d; goal was for restriction but was not seen with intake | Low fat; iso-caloric (per intake not per goal) | Adequate | 0% | 3 |
| Vlachos et al (2011)46 | BMI >30; did include patients receiving insulin (n=79) | NA | “Low-carbohydrate and protein sparing modified fast”; calorically restricted | Low glycemic | NA | 22% | 6 |
| Westman et al (2008)47 | BMI 27-50; did include patients receiving insulin (n=97) | 52/52; 78% female | <20 g/d; intense behavioral support | Low glycemic; iso-caloric (per intake not per goal) | Adequate | 48% | 6 |
| Zadeh et al (2018)48 | BMI 30-38; did include patients receiving insulin (n=42) | 46.5 | 20%; intense behavioral support; calorically restricted | Low fat; high fat; moderate fat; probably iso-caloric (three control arms) | NA | 7% | 6 |
| Daly et al (2006)49 | Obese, poorly controlled T2D; did include patients receiving insulin (n=102) | 58/59; 52% female | <7 0 g/d | Low fat; iso-caloric | Adequate | 23% | 3 |
| Davis et al (2009)50 | BMI>25; A1C 6-11%; did include patients receiving insulin (n=105) | 54/53; 50% female | 20-25 g/d × 2 weeks + 5 g/wk; calorically restricted | Low fat; iso-caloric | Adherent at VLCD level | 20% | 12 |
| Yancy et al (2010)51 | BMI 27-30 plus obesity related disease, or BMI≥30; did include patients receiving insulin (n=46) | 57/55; 13% female | ≤20 g/d; intense behavioral support | Low fat + orlistat | Adequate | 11% | 12 |
| Samaha et al (2003)52 | BMI >35; did include patients receiving insulin (n=52) | NA | <30 g/d; intense behavioral support | Low fat | NA | 42% | 12 |
| Shai et al (2008)53 | BMI≥27 or coronary heart disease; did include patients receiving insulin (n=46) | NA | 20 g/d × 2 months with gradual increase to maximum 120 g/d | Low fat; Mediterranean (two control arms) | NA | 22% | 24 |
| Lee et al (2013)54 | BMI 30-65 (n=105) | NA | <70 g/d | Low fat | NA | 25% | 6 |
| Breukelman et al (2019)55 | BMI 38.9 (mean) (n=39) | 55/58; 60% female | <50 g/d | No treatment control | NA | 10% | 4 |
| Perna et al (2019)56 | BMI 24.9-34.9; A1c≤7.5% taking metformin; did not include patients receiving insulin (n=17) | 59.5/67.8; 65% female | <125 g/d; calorically restricted | “Standard” calorically restricted + metformin | NA | 0% | 3 |
ADA=American Diabetes Association; BMI=body mass index; HbA1c=glycated hemoglobin; NA=not available; NAFLD=non-alcoholic fatty liver disease; T2D=type 2 diabetes; VLCD=very low calorie diet.
*
Saslow 2014 met inclusion criteria but included 4/34 randomized participants who had pre-diabetes, not type 2 diabetes (>88% diabetic population). Contact with authors for diabetes specific data was unsuccessful. Although this situation was not considered a priori, research team decided to include this study on basis that any study with >80% of population having diabetes would be eligible; this scenario was not relevant in any other cases. This decision was made before results of study were reviewed.