The study, published in the August edition of the European Journal of Cancer Prevention, found users of one type of first generation analogue mobile phone were 30% more likely to develop a brain tumour on average. However those using the Nordic Mobile Telephone (NMT) phones for 10 years or more increased their risk by 80% compared with a control group.

Researchers found that the tumours generally corresponded to the area of the head most exposed to mobile phone radiation, the area behind the ear.

Research team leader Professor Lennart Hardell from the oncology department of Orebro University hospital, said: "We don't think analogue phones are safe. The results are very clear."

Mobile phone manufacturers denied the researchers' findings. A spokesman for Ericsson in Sweden told the Reuters news agency: "The study and the conclusions it reaches differs from at least three other studies in the past in several highly regarded scientific journals. None of these studies found a connection between mobile phones and cancer."

Whether or not mobile phones are involved, the odds of developing a brain tumour are still low. In a population of eight million people in Sweden, Prof Hardell said there are 1,200 reported brain tumours a year.

In addition, most of us now use next-generation digital GSM mobile phones, although NMT phones are still produced. Across western Europe, analogue use is unusual, according to John Delaney, principle telecoms analyst at Ovum.

"You'd have to work hard to buy an analogue phone in Europe," he said, noting that Norway's Telenor shut down its NMT network last year.

Researchers found no correlation between digital phones and brain tumours. Prof Hardell put forward a theory that we simply have not been using the digital phones, in widespread use only since the mid-90s, for enough time for tumours to develop. NMT phones have been in use for over 20 years.

Prof Hardell and his team had compared the mobile phone usage of 1,617 Swedish patients diagnosed with brain tumours between 1997 and 2000 with a control group that did not suffer from brain tumours.

"The risk is higher primarily in the temporal lobe, on the same side as the phone is used. Here the risk is 2.5 times greater. The type of tumour showing the largest increase, a 3.5 times greater risk, is tumours of the auditory nerve," the study found.

Prof Hardell described his own use of a mobile phone as "very sparse".

"I use it always with a headset and for very brief communications, just to say 'I am here', or something like that," he explained. "I never have long conversations over a mobile phone."

He said he recommends his teenage children use the phone sparingly.

However, Mr Delaney said he uses his GSM phone "with no particular concern".

"I don't use a headset and I hold the phone to my ear," he said. He said from a symptomatic point of view - for example headaches and warming behind the ear - analogue had been much worse than digital, but that neither type of phone had been proven to cause cancer.