"Taking antibiotics for more than two weeks increases your risk of bowel cancer by 73 per cent," reports the Daily Mail.
However, the study it reports on did not look at rates of bowel cancer. What it did find is an increased risk of bowel polyps for women who took antibiotics for two months or more.
Bowel polyps are small growths that develop on the lining of the colon or rectum. The majority of these growths are benign (non-cancerous), though it is estimated that without treatment to remove them, a small minority will turn cancerous.
This study included 16,642 women aged over 60, who'd had a colonoscopy, a test used to screen for bowel cancer in the US (not routinely used for screening in the UK). They were asked to remember how much they'd used antibiotics earlier in life.
Those who said they'd used antibiotics for a period of at least two months from the age of 20 to 60 were more likely to have been diagnosed with a colorectal adenoma (more commonly known as a bowel polyp) during colonoscopy.
Antibiotics kill off some of the diverse bacteria that live in the gut, which may lead to an imbalance of bacteria. It's been suggested that this could make the gut more vulnerable to cancerous growths.
However, the study doesn't prove antibiotics directly cause bowel cancer, or even bowel polyps. If you've been prescribed antibiotics, you shouldn't stop taking them because of this study.
The study was carried out by researchers from Harvard Medical School, Harvard TH Chan School of Public Health, the University of Nebraska and Yale School of Medicine. It was funded by grants from the US National Institutes of Health and various charitable bodies. The study was published in the peer-reviewed medical journal Gut.
BBC News covers the study in a balanced and accurate fashion, and includes useful information on known bowel cancer risk factors. In contrast, the Mail Online used a scaremongering headline, based on the subgroup of women found to be at highest risk. But this figure did not take into account multiple risk factors for bowel cancer, so we do not know if it is accurate. The story also downplays the fact that more than 90% of this type of polyp (adenoma) does not become cancerous.
This was a prospective cohort study in which a large number of women where followed up over a long period of time. The intention was to make links between lifestyle and health outcomes. But this type of study cannot prove that one factor (antibiotic use) causes an outcome (bowel polyps). It can only show that they are linked.
The researchers gave women aged 60 or over questionnaires asking about their lifestyle and health. They were asked about use of antibiotics during their younger life, as well as more recently. They were also asked if they'd had a colonoscopy and been diagnosed with polyps as a result.
After adjusting for potential confounding factors, the researchers looked to see whether women who'd been diagnosed with polyps were more likely to have taken antibiotics for long periods in earlier life.
The women were all taking part in the Nurses' Health Study, a long-running study in the US that started in 1976. Women were asked to fill in questionnaires every two years.
For this study, researchers included only women:
They adjusted their figures to take account of many factors linked to an increased risk of colorectal cancer, including poor diet, increasing age, family history of colorectal cancer, diabetes, body mass index (BMI), smoking and lack of exercise. They also adjusted the results for aspirin and hormone replacement therapy (HRT) which are linked to a reduced risk of colorectal cancer.
Researchers followed up the medical records of women who reported having been diagnosed with a polyp, to see where in the bowel it was found, and whether it was high or low risk in terms of how likely it was to turn cancerous. However, they did not report how many women developed bowel cancer.
They designed the study so the figures would not be affected by certain people having multiple colonoscopies. They looked separately at women's antibiotic use in their 20s and 30s, in their 40s and 50s, and more recently.
Of the 16,642 women in the study, 1,195 (7%) had a polyp diagnosed during a colonoscopy.
Compared to women who'd never taken antibiotics:
The chance of having high or low risk polpys was about the same as having any polyps. The chance of having a polyp in the upper region of the colon (called the proximal region) seemed more strongly linked to antibiotic use than the chances of having a lower colon polyp.
The researchers said their results "provide additional support" for linking antibiotic use to bowel cancer, and that – if the findings are confirmed by other studies – they "suggest the potential need to limit the use of antibiotics".
However, they admit that the bacteria being treated by the antibiotics might also have raised inflammation in the body. Inflammation is another risk factor for cancer, so the problem might be the infection, not the treatment.
Antibiotics, like all drugs, have side effects. We know that they affect the composition of bacteria that live in a healthy gut. This study suggests that might possibly be linked to future development of bowel cancer.
However, there are some major limitations to keep in mind. Bowel polyps are very common, and they're not cancerous. Most people who have them won't know they're there, unless they have a colonoscopy. Some polyps do develop into bowel cancer, but we don't know if any of these women got bowel cancer, or how many of their polyps would have become cancerous if not treated.
It's highly possible that women aged 60 might not remember accurately how often they used antibiotics in their 20s, or for how long. So we can't be sure whether women were over-estimating or under-estimating their antibiotics use.
Observational studies such as this cannot show that one factor directly causes another. As the researchers say, antibiotics are used to treat bacterial infections. Infections cause inflammation, and that's a risk factor for cancer. So the study may have measured the effect of repeated exposure to bacterial infection, rather than use of antibiotics.
The study was only carried out in women, so we don't know if the results apply to men.
The study did a good job of controlling for other potential confounding risk factors, but no study can control for everything. It's possible there are other factors involved which we don't know about.
Antibiotics have been over-used in the past and this study is a reminder that they should only be used when necessary. But it doesn't show that they cause bowel cancer. If you're taking antibiotics prescribed by your doctor to treat an infection, you should carry on taking them. Not doing so could contribute towards the ongoing problem of antibiotic resistance.
Known risk factors for bowel cancer include:
Avoiding these risk factors should help lower your risk of bowel cancer.