Monday January 4 2010
“One in six Britons with high blood sugar levels faces a greater danger of developing cancer,” The Observer reported.
This story is based on research that found an association between high blood glucose levels and an increased risk of certain types of cancer over an average of 10 years of follow-up.
Although the research did find an association between high blood sugar levels and cancer risk, there are many other lifestyle, medical and genetic factors that may contribute to a person’s likelihood of developing cancer, few of which were considered in this analysis. Also, the study can only demonstrate a link between high blood sugar and cancer. It cannot indicate that one causes the other.
Despite these limitations, it is known that healthier diets and lifestyles are associated with a lower risk of cancer. A healthier lifestyle includes limiting sugar intake.
Where did the story come from?
The research was carried out by Dr Tanja Stocks and colleagues from the University of Umeå, Sweden. It was funded by the World Cancer Research Fund and published in the peer-reviewed medical journal PLoS Medicine.
The Observer gave an accurate account of the research, highlighting other factors that can also contribute to cancer risk. However, as little information is given about the various lifestyle factors, the socio-economic status and the ethnicity of the participants of the study, it is not possible to use this cohort to predict the number of people at risk of sugar-related cancer in the British population as a whole.
What kind of research was this?
This prospective cohort study investigated whether there is an association between high glucose levels and the risk of developing cancer. It followed about 275,000 men and 275,000 women from Norway, Sweden and Austria, measuring their glucose levels at the beginning of the study and their risk of cancer during follow-up.
What did the research involve?
The participants were recruited from the Metabolic syndrome and Cancer project, which includes data from populations in Norway, Austria and Sweden. This particular research used pooled data from separate cohorts from each of the three countries.
Participants had not had cancer at the time of the study, and mean age was 44.7 years for men and 45 years for women. Those with extreme metabolic factors, such as very low glucose levels or a BMI of less than 15 or greater than 60, were excluded, as were those who had missing data for smoking status, BMI or glucose levels.
The participants’ height, weight, blood pressure, cholesterol and triglyceride (fat) levels were measured. Blood glucose levels were measured slightly differently in each cohort.
Cancers were classified according to internationally recognised criteria and codes (International Classification of Diseases seventh revision). The site of the cancer was assessed and if more than 50 incidences of cancer were found for each site the relative risks of cancers were calculated separately for men or women. For less common cancer sites, male and female data were combined to calculate the relative risk.
Participants were followed up, and the risk of cancer was calculated from one year after being included in the study to either the date of first cancer diagnosis or cancer death, death through other causes, emigration, or to the end of the cohort study period, whichever occurred first. The average length of follow-up was 11.3 years for men and 9.6 years for women.
The relative risks of developing cancer were categorised according to age and gender. The analyses were adjusted for BMI, age at measurement and smoking status.
The researchers used two methods to test for an association between blood glucose levels and risk of cancer. First they investigated whether there was a change in risk with each increment (1mmol/l) of blood sugar level. Second they compared the risks of individuals in the highest blood-glucose-levels quintile (top fifth) of the cohort with those in the lowest (bottom fifth).
What were the basic results?
In men, there were 18,621 cases of diagnosed cancer and 6,973 cases of fatal cancer in total. With each increment in blood glucose level there was a 5% increase in relative risk of developing cancer and a 15% increase in relative risk of developing fatal cancer (Relative risk (RR) 1.05, 95% confidence interval (CI) 1.01 to 1.10, and RR 1.15, 95% CI 1.07 to 1.22, respectively).
When comparing the highest quintile of blood glucose to the lowest, there was an 18% increase in the relative risk of cancer and a 50% increase in the relative risk of fatal cancer.
In men, significant increases in risk of both cancer and fatal cancer at specific sites per increment in glucose were observed for cancer of the liver, gallbladder and respiratory tract.
The absolute risk of cancer over a 20-year period for a 50-year-old man with blood sugar levels in the lowest 40% and top 10% of the cohort population was 14.0% and 15.7%, respectively, and the corresponding risk of fatal cancer was 5.0% and 8.8%.
In women, there were 11,664 cases of diagnosed cancer and 3,088 cases of fatal cancer in total. With each increment in blood glucose levels there was an 11% increase in relative risk of developing cancer and a 21% increase in the relative risk of fatal cancer (RR 1.11, 95% CI 1.05 to 1.16, and RR 1.21, 95% CI 1.11 to 1.33, respectively).
When the highest quintile of blood sugar was compared to the lowest, there was an increased relative risk of 29% for a diagnosis of cancer and a 69% increase for fatal cancer.
Significant positive associations among women were observed for the incidence of cancer and fatal cancer of the pancreas. A significant association was also observed for the incidence of urinary bladder cancer and for fatal cervix and uterine cancer.
In women, the absolute risks of developing cancer were 12.2% in people with blood glucose levels in the lowest 40% and 16.7% in those with blood glucose levels in the highest 10%, and for cancer death, 3.0% and 6.0%, respectively.
The strength of the association between increased blood glucose levels and the risk of cancer varied between the pooled cohorts, with one female cohort reported as not showing an association.
How did the researchers interpret the results?
The researchers suggest that their study provides “strong evidence that high blood glucose is a risk factor for cancer” and that “associations between glucose and overall incidence and fatal cancer were stronger in women than in men”.
This large prospective cohort study was well conducted. However, there are several facts that should be considered when interpreting the results.
- The study included previous smokers. Although the researchers say that smoking status had no effect on their findings, they also state that imprecise or incorrect categorisation of smoking status may have affected the results.
- The researchers highlight that the different protocols for measuring blood glucose levels may have affected the outcome.
- Information about the participants’ diet and lifestyle was not available, for example exercise levels, alcohol consumption or the many other genetic and medical risk factors that can also contribute to the risk of cancer.
- The 5 and 11% risk increases for men and women had only borderline significance.
Overall, this research provides some evidence that high blood glucose is associated with an increased risk of some types of cancer. However, it did not directly test or suggest that high blood sugar causes these cancers. Although the research did find an association between high blood sugar levels and cancer risk, there are many other lifestyle, medical and genetic factors that may contribute to a person’s likelihood of developing cancer, few of which were considered in this analysis.
It is now known that healthier diets and lifestyles are associated with a lower risk of cancer, and as part of this healthier lifestyle, lower sugar consumption is recommended.