Table 3.

Grading of Evidence for associations between dietary intakes ( ≤ 12 months) and weight loss (up to ten-years) (15 studies)*.

	Macronutrient Composition			Food Pattern
	Carbohydrate	Protein	Fat	Fruit and Vegetables
Study Findings	Inverse Association (3 studies, total N = 1888, up to 10-years) [30–38, 40] Level II: 1 LR study [38] (N = 1610) Level III: 1 LR study [40] (N = 189) Level IV: 1 AR study [30] (N = 89) No Association (5 studies, total N = 368, up to 8-years) [24, 25, 39, 48, 53] Level II: 1 LR study [48] and 2 AR studies [24, 53] (N = 161) Level III: 2 LR studies [25, 39] (N = 207)	Positive Association (5 studies, total N = 2232, up to 10-years) [30, 38, 40, 48, 49] Level II: 3 LR studies [38, 48, 49] (N = 1954) Level III: 1 LR study [40] (N = 189) Level IV: 1 AR study [30] (N = 89) No Association (6 studies, total N = 444, up to 8-years) [24, 25, 27, 29, 39, 53] Level II: 2 AR studies [24, 53] (N = 110) Level III: 2 LR studies [25, 39] and 2 AR studies [27] (N = 334)	Inverse Association (2 studies, total N = 1799, up to 10-years) [38, 40] Level II: 1 LR study [38] (N = 1610) Level III: 1 LR study [40] (N = 189) No Association (5 studies, total N = 400, up to 8-years) [24, 25, 30, 39, 48] Level II: 2 AR studies [24, 48] (N = 101) Level III: 2 LR studies [25, 39] (N = 210) Level IV: 1 AR study [30] (N = 89)	Positive Association (2 studies, total N = 94, up to 1.5 years) [34, 36] Level III: 2 AR studies [34, 36] (N = 94)
GRADING OF EVIDENCE
Evidence Statement	An evidence statement could not be made due to inconsistent evidence.
Evidence Base	C – Satisfactory. Level II to IV (majority being cohort) studies with low to acceptable risk of bias.			C – Satisfactory. Level III (intervention) studies with acceptable risk of bias.
Consistencies	D – Poor. Study findings highly inconsistent. Multiple study designs with varied risk of bias, outcome measures and duration of follow-up.			A – Excellent. All study findings were consistent. Similar study design, risk of bias, outcome measures and duration of follow-up.
Clinical Impact	D – Poor. Inconsistent study findings and design have affected ability to apply to practice. Non-diet moderators to weight change not always addressed.			C – Slight. Though consistent findings, co-variables from study design have prevented attribution of effect to diet intakes alone.
Generalizability	B – Good. All studies in adults at least one-year post-bariatric surgery in an outpatient setting.			C – Slight All studies in adults at least one-year post-bariatric surgery in an outpatient setting. Inclusion of only motivated individuals in one of the study’s interventions.
Applicability	B – Good. Most studies were conducted with population from the Western context like the Australian bariatric context.
RECOMMENDATION				GRADE OF RECOMMENDTION
No recommendations could be drawn. More well designed RCTs or prospective cohort studies may help with clarifying whether and what associations may be present between weight and diet within ≤12 months post-bariatric surgery.				-

*Using NHMRC body of evidence framework [20]. AR Acceptable risk of bias, HR High risk of bias, LR Low risk of bias.