Association between carbohydrate quality and inflammatory markers: systematic review of observational and interventional studies

Background: Chronic low-grade inflammation is a likely intermediary between quality of carbohydrate and chronic disease risk. Objective: We conducted a systematic literature search to evaluate the relevance of carbohydrate quality on inflammatory markers in observational and intervention studies. Design: MEDLINE, EMBASE, and the Cochrane Library were searched for studies on associations between glycemic index (GI), glycemic load (GL), dietary fiber or fiber supplements or whole grain intake, and high-sensitivity C-reactive protein (hsCRP) or interleukin 6 (IL-6). Included studies had to be conducted on adults (healthy, overweight, with type 2 diabetes or metabolic syndrome features, but without inflammatory disease) with >= 20 participants and a 3-wk duration. Results: In total, 22 of the 60 studies that met our inclusion criteria examined GI/GL: 5 of 9 observational studies reported lower concentrations of hsCRP or IL-6 among persons with a lower dietary GI/GL; 3 of 13 intervention studies showed significant antiinflammatory effects of a low-GI/GL diet, and 4 further studies suggested beneficial effects (trends or effects in a subgroup). For fiber intake, 13 of 16 observational studies reported an inverse relation with hsCRP or IL-6, but only 1 of 11 intervention studies showed a significant antiinflammatory effect of fiber intake, and a further trial reported a beneficial trend. For whole-grain intake, 6 of 7 observational studies observed an inverse association with inflammatory markers, but only 1 of 7 intervention studies reported significant antiinflammatory effects, 1 further study was suggestive (in a subgroup) of such, and another study found an adverse effect (trend only). Conclusions: Evidence from intervention studies for antiinflammatory benefits is less consistent for higher-fiber or whole-grain diets than for low-GI/GL diets. Benefits of higher fiber and whole-grain intakes suggested by observational studies may reflect confounding.