Format

Send to

Choose Destination
Eur J Clin Nutr. 2007 Aug;61(8):929-37. Epub 2007 Jan 24.

Limited evidence for effects of diet for type 2 diabetes from systematic reviews.

Author information

1
Department of General Practice, Radboud University Nijmegen Medical Centre, 6500 HB Nijmegen, The Netherlands. f.vandelaar@hag.umcn.nl

Abstract

OBJECTIVE:

Systematic reviews are an appraised method to summarize research in a concise and transparent way, and may enable to draw conclusions beyond the sum of results of individual studies. We assessed the results, quality and external validity of systematic reviews on diet in patients with type 2 diabetes.

DESIGN, SETTING, SUBJECTS:

We systematically searched for systematic reviews on nutritional interventions in patients with type 2 diabetes that used a reproducible search strategy in at least one major database that applied some form of quality assessment. We assessed quality and the external validity of the retrieved systematic reviews. Outcomes were defined as statistical meta-analyses or narrative results using a predefined and reproducible method.

RESULTS:

Six systematic reviews met the inclusion criteria, investigating dietary interventions in general (n=3), chromium supplementation (n=1), fish-oil (n=1) or herbs and nutrition supplements (n=1). Quality assessment showed minimal/minor flaws in four cases and major/extensive flaws in two cases. All reviews had insufficient data needed to judge external validity. In reviews with minimal/minor flaws, we found beneficial effects of very-low-calorie diets and fish-oil supplements. However, the external validity of these results could not be assessed sufficiently.

CONCLUSIONS:

Systematic reviews largely failed to produce knowledge beyond the sum of the original studies. Furthermore, judgment of external validity was hampered in most cases owing to missing data. To improve the quality and usefulness of systematic reviews of dietary interventions, we recommend the application of more focused research questions, but with broader inclusion criteria, for example, the use of observational studies.

SPONSORSHIP:

Internal funding Radboud University Nijmegen Medical Centre.

PMID:
17251927
DOI:
10.1038/sj.ejcn.1602611
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Nature Publishing Group
Loading ...
Support Center