Floodplains of the rivers Rhine and Meuse in the Netherlands are moderately polluted with heavy metals and toxic compounds. In these floodplain areas a number of nature rehabilitation programmes are being planned and executed. The question arises whether or not the contaminants of concern pose a risk to organisms exposed, and, if so, whether this will hinder the expected ecosystem recovery. This study focuses on the effects of contaminants on aquatic and terrestrial macro-invertebrates in the field situation. In three shallow lakes in the floodplains of the river Rhine, showing different degrees of contamination, chemical analyses and laboratory bioassays have been carried out on sediment samples together with intensive biological field surveys (a so called TRIAD-approach). A nearly identical set-up was chosen to assess 18 terrestrial floodplain sites. The concentrations of several contaminants exceed generic critical risk levels. Risk assessment models show that actual effects on top predators cannot be excluded. Species of lower trophic status might be at risk as well. In the aquatic compartments chemical, ecotoxicological and ecological results point in the same direction. They can be summarised as moderate biological effects, which can mainly be attributed to the elevated levels of a mixture of priority pollutants (metals and polyaromatic hydrocarbons). No large differences were found among the three lakes. At terrestrial sites, however, fewer effects were found in field surveys and bioassays, while the chemical analyses revealed the same levels of pollution as in the lakes study. This suggests a lower bioavailability of the contaminants than expected. In conclusion, no extreme effects have been found. This conclusion only partly supports the predictions made by present day-risk assessment models. The advantages of nature rehabilitation, that is, improvement of ecological quality through habitat diversity, seem to be larger than the disadvantages of increasing toxic stress by exposing a larger diversity of species to floodplain contaminants. To confirm this statement rehabilitation programmes should include overall as well as in-depth monitoring studies.