antibiotics increased colon cancer risk
You forwarded this message on Thu 10/17/2019 5:30 PM
Geoff A. Modest, M.D.
Wed 10/9/2019 7:29 AM
A recent analysis found that oral antibiotic use increased the risk of subsequent colorectal cancer (see antibiotic use colon cancer Gutmicrob2019 in dropbox, or doi:10.1136/gutjnl-2019-318593).
Details:
-- Data from the Clinical Practice Research Datalink (CPRD) in the UK, which has the clinical records from population-based prospective data, including 11.3 million patients from 674 practices (6.9% of the UK population)
-- matched case-control study, from 1989-2012
-- median age 72, 45% female, BMI normal 30%/overweight 34%/obese 17%, never smoker 41%/current smoker 5%/ex-smoker 28%, nondrinker 9%/current drinker 25%/ex-drinker 2%, diabetes 8%, chronic aspirin use 19%, colonoscopy 3%
-- people with colorectal cancer (CRC) were more likely to be overweight (35% vs 34%), obese (19% vs 16%), have history of smoking (50% vs 47%), moderate to heavy alcohol use (14% vs 11%), history of diabetes (9% vs 8%), and to have undergone colonoscopy (4% vs 3%)
-- antibiotic use overall (note: these do not add up to 100%, since many patients received multiple classes of medications):
-- penicillins 81%, 93,000 patients
-- macrolides 30%, 35,000 patients
-- trimethoprim/sulfa 29%, 33,000 patients
-- cephalosporins 25%, 29,000 patients
-- quinolones 14%, 16,000 patients
-- tetracyclines 20%, 23,000 patients
-- other 11%, 13,000 patients
--28980 cases of colorectal cancer (CRC) and 137,077 controls, median follow-up 8 years
Results:
-- for any antibiotic use, as compared to 0 days:
-- colon cancer:
--1 to 15 days: 8% increase, 16 to 30 days 14% increase, 31 to 60 days 15% increase, > 60 days 17% increase, with p<0.001 for the trend
-- anti-anaerobic antibiotics: similar numbers, with p<0.001 for the dose-response trend, nonsignificant association for anti-aerobic antibiotics
-- proximal colon: 1 to 15 days: 14%, 16 to 30 days 15%, 31 to 60 days 32%, >60 days not statistically significant 9%, p=0.046 for the trend
-- anti-anaerobic antibiotics: similar numbers, statistically significant increase only in those in the 1-15 or 16-30 day groups for anti-anaerobic antibiotics, but not statistically significant trend for any antibiotics
-- distal colon: no statistically significant relationship, though with anti-anaerobic antibiotics there was an almost statistically significant trend (p=0.056), with a statistically significant 18% increase in those with 16-30 days of antibiotics, HR 1.18 (1.01-1.37)
-- rectal cancer: statistically significant differences: 15% decrease with antibiotics, HR 0.85 (0.79-0.93); 12% decrease in those on >60 days of anaerobic antibiotics; and for the almost statistically significant trends in both antibiotics overall and in the anti-aerobic ones, p=0.054
-- by specific antibiotic classes:
--penicillins, especially ampicillin/amoxicillin increase the risk of colon cancer by 9%, aOR 1.09 (1.05-1.13)
--tetracyclines decreased rectal cancer risk by 10%, OR 0.90 (0.84-0.97)
--limiting analysis to those with at least 15 years of followup: the antibiotic cancer association was found to be 17% increased for antibiotics exposure occurring > 10 years before the diagnosis, aOR 1.17 (1.06-1.31), though no association with rectal cancer risk
--if consider ampicillin/amoxicillin as a primarily anti-aerobic antibiotic, the effects by antibiotic class was reversed (ie, not surprisingly, this group was the dominant contributor)
--overall, there was a non-linear relationship between days of antibiotic use and cancers:
--CRC overall: curvilinear increase up to 60 days, then leveling off
--proximal colon: dramatic increase until 15 days of antibiotic, then gradual decline
--distal colon: slow, smooth increase until 45 days, then leveling off
--rectum: slight increase for 15 days, then swooping decrease for the next 100 days
--and, no difference in those by age <70 vs >70yo
Commentary:
--in 2010, global antibiotic consumption was 70 billion individual doses (10 doses/person), and has been increasing since then
--even narrow-spectrum antibiotics have been associated with strong and persistent effects on the distal gut microbiome, and they seem to impair the integrity of the intestinal barrier, and allow for colonization with pathogenic bacteria
--epidemiologic studies have suggested the relationship between antibiotics and cancer:
--a Finnish cohort study of 3 million people and 134K cancers from the Finnish Cancer Registry found a relationship between antibiotic exposure in 1995-7 and increased risk of prostate (39%), breast (14%), lung (79%) and colon cancers (15%), with increasing cancer risk associated with increasing antibiotic exposure, going from 0-1 to >5 prescriptions
--a Taiwanese case-control study of diabetic patients found a relationship between CRC and anti-anaerobic but not anti-aerobic antibiotics, with OR of 2.31 for colon cancer and 1.69 for rectal cancer (ie, increased risk vs decreased in the UK study above), especially in those with intra-abdominal infections, though there was no dose-response relationship for cumulative use
--a Netherlands study of 4K cancer cases and 16K controls found that increasing antibiotics was associated with a 26% increased risk of CRC (see microbiome antibiotics colon cancer digdissci2016 in dropbox, or DOI 10.1007/s10620-015-3828-0)
--the US Nurses’ Health Study found a relationship between early age antibiotic exposure and high risk of colonic adenomas (see http://gmodestmedblogs.blogspot.com/2017/04/antibiotics-microbiome-changes-and.html )
-- this current UK study found that the increased CRC risk varied by anatomic site (greatest in the proximal colon) and specifically with antibiotics with anti-anaerobic activity. There was an inverse association with rectal cancer and antibiotic use, especially in those with antibiotics used for more than 60 days
--not sure how to explain this: the proximal colon is the part first exposed to the antibiotics, though one study (but not another) did find a decreasing gradient of a specifically bad microbe (Fusobacterium nucleatum, which has been specifically found to be associated with increased CRC risk) over the length of the colon
--in terms of the microbiome changes (and the microbiome is predominantly composed of anaerobes):
--several microbes are potentially carcinogenic: Fusobacterium, Porphyromonas, Enterococcaceae, Bacteroides-Prevotella, and the toxin producing B fragilis and E coli.
--but, there is a caveat: we should not assume that a single specific species is responsible; there might well be interactions between them (and ?others) which really confers the increased risk (ie, it may not be appropriate to focus on a single pathologic microbiome change, since the real clinical outcomes may be related to some complex interaction involving many bacteria)
--and, unfortunately, there are no data available in this study of the specific microbiomes of those who did/did not get CRC
--in terms of length of exposure, it is interesting that overall there is a plateau after about 60 days of exposure. This suggests (to me) that those with longer exposures perhaps have more long-term and/or profound microbiome changes?? and additional antibiotic exposure doesn't really matter??
--the fact that tetracyclines seem to decrease rectal cancer risk. Not clear why. They are anti-inflammatory. ??if that has a role. And, if so, why only for rectal cancer?? Or, is it that the tetracyclines were used to treat different diseases (eg sexually-transmitted ones, which might have a role themselves in rectal cancer)?
-- limitations of the study:
--though this was the largest study (and has really a high-quality, all-inclusive database), and it did include some risk factors for CRC in their adjusted analysis (eg BMI, smoking, alcohol, comorbidity, medications), it did not have information about diet, exercise, or family history, for example, and there is always the potential for residual confounding
--and, one issue which I think dilutes the clinical accuracy is that, though 70% of patients with CRC had been prescribed antibiotics, 68.5% of controls had been (I would welcome comments from those more statistically-savvy than myself… it seems to me that the closer the exposures are in both groups, the more we have to rely on mathematical analysis to sort out associations, and the more prone to less clinically accurate ones???)
--And, 71.3% of patients developing colon cancer had antibiotic use, 68.5% of the control patients had antibiotic use; but, there was no difference in those developing rectal cancer
--for a recent review of some of the studies on antibiotic overprescribing, antibiotic resistance, and effects on the microbiome, see http://gmodestmedblogs.blogspot.com/2019/01/antibiotic-overprescribing-2-more.html
--see http://gmodestmedblogs.blogspot.com/2015/11/longterm-microbiome-changes-with.html, a small study finding microbiome changes for up to one year after a single antibiotic exposure
So, this article adds to the list of articles suggesting that antibiotics lead to significant microbiome changes (that can be long-lasting after even a pretty short antibiotic exposure), predisposing it to the overgrowth of harmful microbes, and, in this case, leading to increased risk of CRC (and other studies suggest increased risk of all kinds of medical problems by microbiome changes resulting from many different types of insults), see http://gmodestmedblogs.blogspot.com/search?q=microbiome for a slew of articles. But a few suggestions from this UK article:
--from their impressive database, they found that only proximal colon cancers were increased; and there are studies suggesting that proximal colon cancers have a significantly worse clinical prognosis than distal ones (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2947632/ , or https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3708260/ ). Of late, there has been a shift to more proximal cancers in the population, with observational studies suggesting that colonoscopy is more effective at preventing left-sided than right-sided (proximal) cancers.
--the article also raises a more specific tilt in how we look at microbiome changes and adverse outcomes from antibiotics, in this case specifically for colon cancer: much of the literature (and my assumption based on some studies looking at specific agents and specific microbiome changes) has been that the broader spectrum antibiotics were necessarily worse because of a more shot-gun disruption of the microbiome. This study found that even pretty narrow-spectrum antibiotics (eg penicillins) may actually be worse than some of the big-hitters, since they have more anaerobic effect (and the microbiome is predominantly composed of anaerobes).
But, the bottom line here: we should really really really limit antibiotic prescribing to the clear antibiotic indications, and (I think) still limit the selected antibiotic to the most specific ones that will work, perhaps taking into account their anaerobic potency. And, from a public health perspective, continue to fight for limiting antibiotic use in livestock to those with clear indications, since that livestock get 70-90% of all antibiotics and they then serve as a huge breeding ground for antibiotic resistant bugs (as per many prior blogs on antibiotic overuse and resistance, including the quite dire warnings of the World Health Organization, eg see http://gmodestmedblogs.blogspot.com/2017/03/antibiotic-resistant-bacteria-of-concern.html) ….
geoff
If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org
to get access to all of the blogs:
2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category
3. or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them
please feel free to circulate this to others. also, if you send me their emails, i can add them to the list
Comments
Post a Comment
if you would like to receive the near-daily emails regularly, please email me at gmodest@uphams.org