Having a persistent build-up of plaque on your teeth could up your risk of dying early from cancer, new research has suggested.

Dental plaque, comprising bacteria that builds up on the surface of teeth, can lead to tooth decay and inflammation of the gums, but now researchers have revealed a potentially more serious consequence of not looking after your teeth properly.

Swedish scientists tracked the health of almost 1,400 randomly selected adults from Stockholm for 24 years up to 2009.

They assessed the oral hygiene of the participants to assess their levels of dental plaque, tartar, gum disease, and tooth loss. They were found to have no obvious gum disease, but substantial levels of plaque were found on the surface of their teeth.

Fifty-eight of the study participants has died by 2009 and 35 of these deaths were caused by cancer. One-third of the deaths were among women and the average overall age of death was around 60.

The researchers pointed out that both the women and men who died could have been expected to live longer than this under normal circumstances. Dental plaque levels in those who had died was higher than in those who lived.

Deaths among the women were mainly caused by breast cancer, while the men died from various types of cancer.

The researchers said after usual potential risk factors for early death were accounted for, the link observed between age, male gender, plaque levels, and premature death were still strong.

Dental plaque was associated with a significantly greater risk of premature death, although the absolute risk of premature death was low - only 58 out of 1,390 participants died after 24 years.

The researchers advise caution in interpreting the figures, which they stress, do not prove conclusively that dental plaque can cause cancer.

"Our study hypothesis was confirmed by the finding that poor (mouth) hygiene, as reflected in the amount of dental plaque, was associated with increased cancer mortality. Further studies are required to determine whether there is any causal element in the observed association."

The research, carried out at the Karolinska Institute in Sweden, is published in the online journal BMJ Open.