The food system is a major source of environmental impact, and dietary change has been recommended as an important and necessary strategy to reduce this impact. However, assessing the environmental performance of diets is complex due to the many types of foods eaten and the diversity of agricultural production systems and local environmental settings. To assess the state of science and identify knowledge gaps, an integrative review of the broad topic of environment and diet was undertaken, with particular focus on the completeness of coverage of environmental concerns and the metrics used. Compared with the 14 discrete environmental areas of concern identified in the United Nations Sustainable Development Goals, the located journal literature mainly addressed greenhouse gas (GHG) emissions and, to a lesser extent, land and water use. Some relevant concerns were rarely addressed or not addressed at all. In the case of GHG emissions, changes in land use and soil carbon stocks were seldom considered. This represents a disconnect between the science informing strategic climate action in the agricultural sector and the science informing public health nutrition. In the case of land and water use, few studies used metrics that are appropriate in a life-cycle context. Some metrics produce inherently biased results, which misinform about environmental impact. The limited evidence generally points to recommended diets having lower environmental impacts than typical diets, although not in every case. This is largely explained by the overconsumption of food energy associated with average diets, which is also a major driver of obesity. A shared-knowledge framework is identified as being needed to guide future research on this topic. Until the evidence base becomes more complete, commentators on sustainable diets should not be quick to assume that a dietary strategy to reduce overall environmental impact can be readily defined or recommended.