Overall, Bt-modified crops appear to be environmentally safe. The proteins produced by Bt have been used in sprays for agricultural weed control in France since 1938 and the USA since 1958 with seemingly no ill effects on the environment.
Bt toxins are considered to be environmentally friendly by many farmers and may be a potential alternative to broad-spectrum insecticides. The toxicity of each Bt type is limited to one or two insect orders; it is nontoxic to vertebrates and many beneficial arthropods, because Bt works by binding to the appropriate receptor on the surface of midgut epithelial cells. Any organism that lacks the appropriate receptors in its gut cannot be affected by Bt.
There is clear evidence from laboratory settings that Bt toxins can affect nontarget organisms. Usually, but not always, affected organisms are closely related to intended targets. Typically, exposure occurs through the consumption of plant parts, such as pollen or plant debris, or through Bt ingestion by their predatory food choices. The methodology used by these researchers has been called into question. Due to significant data gaps, the real-world consequences of Bt transgenics remains unclear.
The observed changes have been found to be of no biological significance by the European Food Safety Authority.
Constant exposure to a toxin creates evolutionary pressure for pests resistant to that toxin. Already, a diamondback moth population is known to have acquired resistance to Bt in spray form (i.e., not engineered) when used in organic agriculture. The same researcher has now reported the first documented case of pest resistance to biotech cotton.
One method of reducing resistance is the creation of non-Bt crop refuges to allow some nonresistant insects to survive and maintain a susceptible population. This technique is based on the assumption that resistance genes will be recessive.
This means that with sufficiently high levels of transgene expression, nearly all of the heterozygotes,i.e. the largest segment of the pest population carrying a resistance allele, will be killed before they reach maturity, thus preventing transmission of the resistance gene to their progeny. The planting of refuges (i. e., fields of nontransgenic plants) adjacent to fields of transgenic plants increases the likelihood that homozygous resistant (s/s) individuals and any surviving heterozygotes will mate with susceptible individuals from the refuge, instead of with other individuals carrying the resistance allele. As a result, the resistance gene frequency in the population would remain low.
Nevertheless, limitations can affect the success of the high-dose/refuge strategy. For example, expression of the Bt gene can vary. For instance, if the temperature is not ideal, this stress can lower the toxin production and make the plant more susceptible. More importantly, reduced late-season expression of toxin has been documented, possibly resulting from DNA methylation of the promoter. So, while the high-dose/refuge strategy has been successful at prolonging the durability of Bt crops, this success has also had much to do with key factors independent of management strategy, including low initial resistance allele frequencies, fitness costs associated with resistance, and the abundance of non-Bt host plants that have supplemented the refuges planted as part of the resistance management strategy.