The paper delves into a key aspect of so-called Bt corn and cotton -- plants that carry a gene to make them exude a bacterium called Bacillus thuringiensis, which is toxic to insects.
Publishing in the journal Nature Biotechnology, US and French researchers analysed the findings of 77 studies from eight countries on five continents that reported on data from field monitors.
Of 13 major pest species examined, five were resistant by 2011, compared with only one in 2005, they found. The benchmark was resistance among more than 50 per cent of insects in a location.
Three of the five cases of resistance were in the US, which accounts for roughly half of Bt crop plantings, while the others were in South Africa and India.
The authors said they picked up a case of early resistance, with less than 50 per cent of insects, in yet another US cotton pest.
And there were "early warning" signs (one percent resistance or less) from four other cotton or corn pests in China, the US and the Philippines.
The scientists found big differences in the speed at which Bt resistance developed.
What made the difference was whether farmers set aside sufficient "refuges" of land for non-BT crops, said the study's authors.
The idea behind such refuges comes from evolutionary biology.
The genes that confer resistance are recessive, meaning that insects can survive on Bt plants only if they have two copies of a resistance gene -- one from each parent.
Planting refuges near Bt crops reduces the chances of two resistant insects mating and conferring the double gene to their offspring.
Practical evidence of this is shown in the case of a cotton-munching pest called the pink bollworm, said his colleague, Bruce Tabashnik.
Bt crops in the southwestern United States, where growers work closely with scientists to devise a refuge strategy, do not have a resistance problem.
In India, though, local pink bollworms became resistance within six years, simply because farmers did not follow the guidelines or get this support.
