Consuming two diet drinks a day could increase the risk of early death by more than a quarter, the World Health Organisation has warned.
The global study of more than 450,000 adults in 10 countries - including the UK - found that daily consumption of all types of soft drinks was linked with a higher chance of dying young.
But the rates for those drinking artificially-sweetened beverages were significantly higher than those consuming full sugar versions, the WHO research found.
>
Their experts today said consumers were better off sticking with water.
The research, which tracked participants for an average of 16 years, is the largest study to examine links between soft drink consumption and mortality.
The new research found death rates among those consuming at least two diet drinks a day were 26 percent higher than among those who had less than one month.
This group also saw their chance of being killed by cardiovascular disease rise by 52 percent.
The findings suggest UK Government policies aimed at cutting sugar consumption - such as the sugar tax on fizzy drinks, and "reformulation" of common sweet foods - could have disastrous consequences.
The Ministry of Health in New Zealand is currently looking into a sugar tax but the Government continues to rule one out for the time being.
Mortality rates were also higher among those regularly drinking sugary drinks.
But overall, they were only eight percent more among those drinking two such drinks a day, compared with those having less than one month.
The findings were published in JAMA Internal Medicine.
Experts said it was possible that people drinking diet drinks were doing so because they were obese or had diseases such as diabetes, but said the study had tried to adjust for that.
Professor Mitchell Elkind, incoming president of the American Heart Association, said "other studies have suggested biological mechanisms may include an impact on insulin signalling in the liver.
The study, led by the International Agency for Research on Cancer, part of WHO, was observational - meaning it did not prove that the drinking habits caused the higher death risks.