O'Connor 2013.

Methods	Randomised controlled trial (parallel design)
Participants	Setting: laboratory; USA n = 40 untrained men (20) and women (20); age range 18 to 39 Placebo group (n = 20; 10 women) Antioxidant group (n = 2; 9 women) Inclusion/exclusion criteria Excluded were pregnant women, individuals engaged in vigorous activities on a regular basis (more than once per week on average during the prior month), those with contraindications to exercise, those taking pain or prescription drugs (except for oral contraceptives), and high consumers of 55 fruits, vegetables, drinks or chocolate products known to contain polyphenols (> 1 serving per day during the prior month)
Interventions	Intervention Grape beverage ‐ stilbenes (resveratrol = 1.75 mg/kg), catechins (catechin = 19.7 mg/kg, epicatechin = 12.6 mg/kg), anthocyanins (peonidin = 31.7 mg/kg, cyanidin = 125 mg/kg, malvidin = 145.2 mg/kg) and flavanols (quercetin = 32.6 mg/kg, kaempferol = 5.6 mg/kg, isorhamnetin = 6.8 mg/kg) Placebo A combination of starch, sugars, cellulose, acids, dipotassium phosphate, potassium citrate, grape flavouring and food dyes was used to create a placebo powder that matched the composition and caloric content of the grape powder as closely as possible but without any polyphenolic compounds Duration 45 days
Outcomes	PRIMARY Delayed onset muscle soreness in the arm was rated in response to isometric strength measurement trials using a 100 mm visual analogue scale where 0 is "no pain" and 100 is "extreme pain" SECONDARY Range of motion Isometric strength of the elbow flexors at 90 degrees Maximal oxygen consumption performed on a treadmill
Exercise type	3 sets of 6 repetitions were performed using the non‐dominant elbow flexors with a weight of 120% of concentric 1 RM
Sources of funding	Financial support for the study was provided by the California table Grape Commission
Notes	One serving of powder daily mixed with 236 mL of water
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	One investigator generated an allocation sequence using an online random number generator that created blocks of 2 with an allocation ratio of 1:1 to the conditions. A different investigator randomly assigned participants in blocks to their groups.
Allocation concealment (selection bias)	Low risk	Opaque containers and a clip was placed on the nose
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind: participants as well as researchers administering the intervention, assessing the outcomes and conducting the statistical analyses were blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study with no adverse effects
Selective reporting (reporting bias)	Low risk	No published protocol available All outcomes reported at all time points Manuscript states: "No adverse events occurred in response to beverage consumption"
Other bias	Low risk	Participants were asked to record food frequency and were asked to refrain from using anti‐inflammatory drugs or other supplements that could reduce soreness