For at least a decade, evidence has been accumulating about the association between antibiotic use and increased risk for colorectal cancer (CRC). Increasingly, it appears that bugs in the gut may play a role.
For over 20 years, Cynthia Sears, MD, has been studying how bacteria contribute to the development of colon cancer. She is a molecular biologist and Professor of Medicine at the Johns Hopkins Kimmel Cancer Center. In addition, she is part of a large project studying the role of the microbiome in colon cancer, which is funded by Cancer Research UK. Goals of the project include identifying a set of bacteria associated with the likelihood of developing adenomas, and finding biomarkers to better identify early-stage colon cancer.
Sears first started her career by studying a single type of bacteria, Bacteroides fragilis. While considered a commensal organism, strains of B. fragilis can produce toxins that cause diarrheal disease and gut dysbiosis (Curr Opin Infect Dis 2010; doi: 10.1097/QCO.0b013e32833da1eb).
When Sears and her team studied B. fragilis in the lab, they kept bumping into cancer pathways. Using a mouse model susceptible to colon tumors, they found that mice colonized with B. fragilis developed excessive numbers of colon tumors. Over time, they transitioned from studying a single bug into looking at the entire microbial community in the colon.
That lead to a headline-grabbing study published in the journal Gut in 2019 (doi: 10.1136/gutjnl-2019-319792). It was the first population-based study to look at the link between antibiotic exposure and risk of CRC, and used data from the Clinical Practice Research Datalink (CPRD). Containing information on over 11 million patients in the U.K., the CPRD represents one of the world's largest medical records databases.
The analysis covered 23 years (from 1989 to 2012), and included records on 28,890 patients with CRC, matched by age, gender, and geography to 137,088 healthy controls. Results showed a dose-dependent association between use of oral antibiotics and increased risk of colon cancer (ptrend <0.001). Risk was greatest in the proximal colon (colon versus rectum, pinteraction<0.001; proximal colon versus distal colon, pinteraction=0.019), and with antibiotics that have anti-anaerobic activity, such as penicillins (aOR=1.09 (1.05-1.13)).
"There was an association between antibiotic exposure and subsequent colon cancer, but it was only in the right colon," said Sears. "Penicillins appeared to track to the increase in right colon tumors."
Furthermore, antibiotic use was associated with colon cancer only when the antibiotics were taken 10 or more years before diagnosis, not sooner.
"That actually makes sense because colon cancer is a slow, chronic disease. It takes about 10 years from the time a cell becomes cancerous until it expands into a colon tumor and is diagnosed," said Sears.
It also addresses the issue of confounding by association, and provides stronger evidence that the association may be real. For example, when patients are already experiencing symptoms of colon cancer, they may be more likely to receive antibiotics, creating a false association, she explained.
Unlike past studies that did not control for other risk factors associated with colon cancer (such as obesity, smoking, alcohol use, and diabetes), Sears and colleagues adjusted for multiple confounders, and the relationship still held.
Other large studies have found similar relationships. A recent analysis included data from 16,642 women aged 60 and over who were participants in the Nurse's Health Study. Results showed that increased duration of antibiotics at ages 20-39 (ptrend =0.002) and 40-59 (ptrend =0.001) was significantly associated with increased risk for colorectal adenoma, a precursor of most CRC. Antibiotic use within the past 4 years was not linked to risk for adenoma (ptrend =0.44). The study similarly controlled for multiple variables associated with risk for CRC (Gut 2018; doi: 10.1136/gutjnl-2016-313413).
Another large case-control study in the U.K. included data from 35,214 patients diagnosed with CRC between 2008 and 2018, matched with 60,348 controls. Results showed a dose-response relationship between CRC and antibiotic use, with higher risk among patients who took antibiotics up to 15 years before diagnosis (OR, 1.90, 95% CI, 1.61-2.19, p < 0.001). Results were also adjusted for multiple confounders associated with risk for CRC (Br J Cancer 2020; doi: 10.1038/s41416-019-0701-5).
Identifying the Mechanism
Sears emphasized that her team's study cannot prove causality, but she also said that she would probably consider antibiotic use a risk factor for colon cancer.
"Without a doubt, every time you take an oral antibiotic, you disrupt the bacteria in your colon. That appears to have consequences in creating a microbiota that is not healthy," she said.
Other researchers have observed gut dysbiosis in CRC. The pattern shows decreased bacterial diversity, along with decreases in butyrate producing commensals such as Bifidobacterium and Clostridia, and increases in pro-inflammatory bacteria such as Fusobacterium nucleatum and Bacteroides fragilis. In general, the hypothesis is that antibiotic use may lead to dysbiosis, predisposing to inflammation, initiation, and progression of CRC (Cancers 2020; doi: 10.3390/cancers12061406).
Sears said that results from her study showing an increased risk for tumors in the right colon makes sense from this standpoint. Oral antibiotics are not completely absorbed, and are found at higher concentrations when they start transiting through the gut in the right colon.
"It's an interesting hypothesis that the drug would hit the right colon first and potentially create more dysbiosis than in the far distal colon," she said. "It's well-known that antibiotics in the lumen of the gut are in higher concentration and can have much more extensive activity than they were designed to have. They're right there with the bugs."
She added that the results showing a link between penicillins and cancer in the right colon also make sense, and probably the entire class of penicillins are involved. Penicillins are active against anaerobes, and they kill off many of the bacteria in the gut that thrive in low oxygen environments. But some gut bacteria-particularly B. fragilis-are naturally resistant to penicillin. So wiping out the "good" bacteria with a course of penicillin leaves bacteria like B. fragilis behind, which could contribute to dysbiosis and eventually CRC.
Big Picture: Judicious Use of Antibiotics
Beyond gut bugs, Sears says an overarching message of her research is the need for more judicious use of antibiotics. She emphasizes careful use of antibiotics, both to preserve them for treating infections as well as to prevent health problems potentially caused by exposure to antibiotics over time, such as CRC.
"Antibiotics are very important drugs but we're overusing them," she said. "What we have on a global basis is a crisis of antimicrobial resistance where we are actually facing infections now that we cannot treat. The big picture message is the need for judicious use of antibiotics."
Veronica Hackethal is a contributing writer.