Is light-at-night a health risk factor or a health risk predictor?

In 2007, the IARC (WHO) has classified "shift-work that involves circadian disruption" as potentially carcinogenic. Ample evidence leaves no doubt that shift-work is detrimental for health, but the mechanisms behind this effect are not well understood. The hormone melatonin is often considered to be a causal link between night shift and tumor development. The underlying "light-at-night" (LAN) hypothesis is based on the following chain of arguments: melatonin is a hormone produced under the control of the circadian clock at night, and its synthesis can be suppressed by light; as an indolamine, it potentially acts as a scavenger of oxygen radicals, which in turn can damage DNA, which in turn can cause cancer. Although there is no experimental evidence that LAN is at the basis of increased cancer rates in shiftworkers, the scenario "light at night can cause cancer" influences research, medicine, the lighting industry and (via the media) also the general public, well beyond shiftwork. It is even suggested that baby-lights, TVs, computers, streetlights, moonlight, emergency lights, or any so-called "light pollution" by urban developments cause cancer via the mechanisms proposed by the LAN hypothesis. Our commentary addresses the growing concern surrounding light pollution. We revisit the arguments of the LAN theory and put them into perspective regarding circadian physiology, physical likelihood (e.g., what intensities reach the retina), and potential risks, specifically in non-shiftworkers.