Colon cancer: increased with antibiotic use

 Colon carncer: increased with antibiotic use

Geoff A. Modest, M.D.
Mon 11/8/2021 8:04 AM
  •  Geoff A. Modest, M.D.
 A recent large data-mining study from Sweden found an increase in colorectal cancer (CRC) incidence with increasing prior antibiotic use (see colon ca antibiotic use JNCI2021 in dropbox, or doi: 10.1093/jnci/djab125)

Details:  

-- A Swedish nationwide complete population-based study with a matched case-control design (matching each colorectal cancer case with 5 matched, cancer-free controls), with information from July 1, 2005 (the start of their registry) to December 31, 2016

    -- 40,545 CRC cases were compared to 202,720 controls

-- 53% men, education 37% primary school/40% secondary school/21% post-secondary school, 53% married/19% widowed/12% unmarried/16% divorced, specialist visits (they consider a surrogate for comorbidities) 27%, hospitalizations within the study period 5%/2 years before the case diagnosis 2%

-- antibiotic exposure: no use 20% (18.7% in cases and 22.4% in controls), low use (1 to 10 days) 14%, moderate (11 to 60 days) 45%, high (61 to 180 days) 60%, very high (>180 days) 4%

-- mean age at CRC diagnosis: 72 years

-- colon cancer case sites: 36% proximal colon, 39% distal, 33% rectal cancer

-- median follow-up: time 8 years

    -- they excluded cancer cases identified within two years of diagnosis (in order to exclude reverse causation: people with cancer needing antibiotics), which means that the median follow-up time was actually 6 years

-- main results: relationship between outpatient antibiotic prescriptions filled and the later development of CRC, including by site of cancer, intensity of antibiotic use, and specifics of antibiotics prescribed

-- the odds ratios below were adjusted for: age, sex, county, levels of eduction, country of birth, marital staus, number of sepcialist visits and hospitalizations

 

Results:

overall CRC risk (for the full 8 year follow-up):

    -- moderate antibiotic use versus no use: 15% increase, odds ratio 1.15 (1.12-1.18)

    -- very high antibiotic use versus no use: 17% increase , OR 1.17 (1.10-1.24)

-- by tumor site (excluding the past 2 years, as noted above):

    -- proximal colon cancer:

        -- moderate use, 9% increased risk, OR 1.09 (1.05-1.14), vs no use

        -- very high use, 17% increase, OR 1.17 (1.05-1.13), vs no use

        -- for the trend of increasing use, p<0.001

    -- distal colon cancer: no significant difference by antibiotic use

    -- rectal cancer:

        -- high antibiotic use: 9% decreased risk, OR 0.91 (0.84- 0.97) 

        -- very high use, 9% decrease, OR 0.91 (0.80-1.04)

        -- for the trend, p<0.001

            --further stratification by sex:

                -- significant trend was only observed in women:

                    -- moderate antibiotic use, 14% decrease, OR 0.86 (0.80-0.92)

                    -- high use: 16% decrease, OR 0.84 (0.276 0.94)

    

-- Association with nine different classes of antibiotics:

    -- overall colorectal cancer: increased CRC risk in men with quinolones or sulfonamides and/or trimethoprim; not much change in women

    -- proximal colon cancer: increased risk in men with quinolones, sulfonamides and/or trimethoprim, and anti-aerobic antibiotics; worse in women with tetracyclines, quinolones, sulfonamides/trimethoprim, and metronidazole/tinidazole

    -- distal colon cancer: worse in men with sulfonamides and/or trimethoprim; better in women with macrolides/lincosamides, metronidazole/tinidazole

    -- rectal cancer: worse in men with sulfonamides and/or trimethoprim; better in women with narrow or broad-spectrum antibiotics (aerobic and anaerobic), quinolones, nitrofurantoin, sulfonamides and/or trimethoprim, macrolides and/or lincosamides,and  metronidazole/tinidazole

-- methenamine hippurate (a urinary tract antiseptic that does not affect the gut microbiota, used to prevent or control recurrent UTIs): no effect on CRC

Commentary: 

-- several studies have found a relationship between antibiotic use and colorectal cancer, this study adding to the litany

    -- for example, a recent study from the UK (see colon ca antibiotic use UK gut2019 in dropbox, or doi.org/10.1136/ gutjnl-2019-318593) also found a gradient of risk by the amount of antibiotics taken, more CRC in the proximal colon associated with antibiotics having more anti-anaerobic activity (especially amoxicillin), and reduced risk of rectal cancer (especially with tetracyclines), specifically with antibiotic exposure more than 10 years before diagnosis

-- The Nurses' Health Study confirmed the relationship between antibiotic exposure and colorectal adenomas, presumably the antecedents to cancer (see http://gmodestmedblogs.blogspot.com/2017/04/antibiotics-microbiome-changes-and.html )

-- it was notable that there was no effect of methenamine hippurate on CRC cancer risk, which is consistent with the hypothesis that microbiota changes are the true association, given that methenamine hippurate does not affect the microbiome

-- research has suggested that certain anaerobic microbes such as Fusobacterium nucleatum are implicated in CRC development, and that the concentration of this organism is highest in the proximal colon and decreases gradually from the cecum to the rectum with antibiotics

-- another recent article reviewed the role of nontoxigenic Bacteroides fragilis, noting that this bacterium is increased in precancerous colonic polyps, has lots of genes involved in lipopolysaccharide (LPS) biosynthesis, stimulates pro-inflammatory cytokine synthesis as well as further LPS biosynthesis within the polyps, and likely increases activation of the immune system and increased inflammation, which may  lead to  carcinogenesis (see colon ca gut microbiome cellhostmicro2021 in dropbox, or doi.org/10.1016/j.chom.2021.08.013)
-- and another recent article reviewed the data on bile acids (bad for microbiome) and short-chain fatty acids (seem to help the microbiome): see microbiome short-chain fatty acids intjmolecsci2019 in dropbox, or doi:10.3390/ijms20051214 ). 

    -- so, the purported mechanism is that "the observed risk gradient along the colorectal continuum is consistent with a high microbial impact in the proximal colon and a deceasing concentration of short-chain fatty acids along the colon, resulting in higher bacterial activity, biofilm formation, and fermentation in the proximal compared with the distal colon and rectum" 

-- the increasing risk of colon cancer with increasing antibiotic exposure reinforces the importance of minimizing antibiotic exposure, which thereby increase the likelihood of anaerobic bacteria proliferation (e.g. Fusobacteria and Bacteroides) which may contribute to CRC development [of note, the UK study above found higher risk with anaerobic antibiotics, vs the current Swedish study which did not really find much difference between aerobic vs anaerobic antibiotics.  ??may not matter which types of microbiome-disruptive antibiotics are used; the issue is  microbiome dysbiosis that leads to overgrowth of bad bacteria with any dramatic microbiome changes...]

-- many of the established CRC risk factors, including excess body fat, dietary factors, lack of exercise, and smoking also alter the gut microbiota

    -- though antibiotics have a more profound effect on the microbiota, including the intestinal overgrowth of such bad actors as Clostridium difficile by depleting some microbiota and providing a hospitable place for these pathogenic bacteria

-- one advantage of these types of studies as done in Sweden is that, unlike the United States, they have a large database linking medical data, prescription drug data, etc. These types of massive and inclusive studies allow much more insight into epidemiology and potential causation. More than 98% of all diagnosed CRC cases are in the Swedish registry, along with links to prescription drugs, medical records, etc.

Limitations:

-- the data from the study began in 2005, and it is quite likely that earlier exposure to antibiotics may be even more important given the likely 10+ year history of the typical progression of colon cancer. In this study, the follow-up time was between 2.5 and 9.5 years (median of 6 years, since they excluded the last two years). indeed, the above-cited UK study found increases in C RC after 10 years. This may be the reason that the proximal colon cancers found that were associated with antibiotics tended to be in the earlier stages of development (stage I or II). and longer followup in sweden might find many more CRCs...

-- we do not have any granular information on the presence of rectal chlamydia. The authors comment that the decreased risk of rectal cancer in women on antibiotics may be because women tend to have more rectal chlamydial infections, chlamydia infections may have a malignant potential presumably by triggering inflammation and reducing apoptosis, and the use of antibiotics may actually treat the chlamydia and thereby lower rectal cancer risk. studies have suggested that women make up 57% of all rectal chlamydia cases detected and in 1/2 of the cases there is no history of anal sex (suggesting that the close proximity of the vaginal and anal orifices in women may be lead to rectal colonization by chlamydia. and perhaps other microorganisms of concern???), see https://clinicaltrials.gov/ct2/show/NCT04030949 

-- no granular information on other causes of microbiome changes, such as excess body fat, diet, exercise, smoking. Though it is likely that antibiotics create the most profound microbiome changes. Also no history of comorbidities (e.g. inflammatory bowel disease) that might have altered CRC risk

-- antibiotic prescriptions were only for outpatient antibiotics (and much more powerful ones might be given in-patient). also unclear how they classified people given more than one different antibiotic over the course of the study

-- there may have been increased CRC screening programs and increased CRC detection (ie, detection bias, which might influence the results), though they do comment that there was no national CRC screening in Sweden during this time period of the study

-- information available to the researchers did not include medication adherence, only prescriptions filled

So, yet another of a long series of articles (and blogs) on the potential effect of adverse microbiome changes on the development of disease. This is one of several studies that have shown increased incidence of colonic adenomas and colorectal cancer with antibiotic use. All of this adds to the imperative for us and patients to minimize antibiotic use, of which a very large percentage are not in fact clinically indicated (see below). This also calls for the importance of decreasing antibiotic use in animals (which still get a large percentage of antibiotics produced and then secondary effects on us through selection of resistant bugs and need for very strong, and even more microbiome-changing antibiotics), and the importance of continued antibiotic development for the increasing strains of antibiotic-resistant bacteria, some resistant to all of the antibiotics we have (eg see http://gmodestmedblogs.blogspot.com/2018/06/s-typhi-and-shigella-resistance.html for more on the huge issue of many increasing antibiotic-resistant bugs, but the dreadful lack of development of new antibiotics)

Interestingly, there was a new study scanning 22.3 million antibiotic prescriptions in ambulatory setting in the US for 8.6 million enrollees in private health plans (Optum Clinformatics Data Mart Database), from 2016-18 (see antibiotic overprescribing non-visit based openforumID2021 in dropbox, or doi.org/10.1093/ofid/ofab412 ), finding that non-visit based antibiotic prescriptions were:

--31% of all antibiotic fills

--lower for children than adults, increasing to 34% in those >65yo, highest for penicillins (36%), and lowest for cephalosporins (25%) and macrolides (25%) and higher for specialist physicians (38%), then internists (28%), family medicine (20%) and pediatricians (10%). overall more than half of ambulatory antibiotic use was either non-visit-based or visit-based but non-infection-related (per relevant infection-related ICD-10 code)

--only 47% of the antibiotic fills were associated with a clinician visit and infection-related diagnosis!!!!

--of course, this is gross data-mining without lots of granular information. for example, it is likely reasonable to prescribe antibiotics over the phone for some conditions, such as a woman with their recurrent usual symptoms of a urinary tract infection that respond to antibiotics and no recent UTI (ie not a resistant infection), no significant fever, etc.  Would be good to know what conditions were being treated (eg acute bronchitis or URIs do NOT have an antibiotic indication)
--but, nonetheless, pretty striking data...... 

-- for an array of blogs on antibiotic overprescribing, see http://gmodestmedblogs.blogspot.com/2019/01/antibiotic-overprescribing-2-more.html 

-- for an array of blogs on the microbiome, including mechanism by which microbiome changes might lead to a variety of diseases including NAFLD, cancer, diabetes, metabolic syndrome, heart disease….  see http://gmodestmedblogs.blogspot.com/search/label/microbiome   ​ 

-- for another array of blogs, but dealing with the consequences of overuse of antibiotics in humans and livestock, and microbial resistance See http://gmodestmedblogs.blogspot.com/search/label/antibiotic%20resistance ​  

-- for some blogs on successful antibiotic stewardship programs, see http://gmodestmedblogs.blogspot.com/2021/03/antibiotic-stewardship-and.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 options):

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

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

 

or: go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs (though there have been a few that did not upload over the last year or two). but overall it is much easier to view blogs and displays more at a time.

 

 

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

Popular posts from this blog

cystatin c: better predictor of bad outcomes than creatinine

Image
  another article found significant discordance between creatinine and cystatin C in assessing renal function,  finding that the latter correlated most strongly with both renal and non-renal adverse clinical outcomes (see   ckd cystatin c better than creat AJKD2023  in  dropbox or doi: 10.1053/ j.ajkd.2023.04.002)   Details: -- 158,601 adults who were in the Stockholm Creatinine Measurements (SCREAM) project and had testing of creatinine and cystatin C on the same day during a healthcare encounter, from 2010-2018     -- >400,000 outpatient tests were done for both creatinine and cystatin C levels -- mean age 62, 48% female -- mean eGFR by creatinine (   eGFR cr ) 80; mean eGFR by cystatin C ( eGFR cys  ) 73 -- mean stage of renal insufficiency: 39%  eGFR cr   90+; 38%  eGFR cr  60-89; 12%  eGFR cr  45-59; 7.1%  eGFR cr  30-44; 3.2%  eGFR cr  15-29, 0.7%  eGFR cr  <15 -- mean urinary albumin-to-creat ratio: 12 mg/g -- comorbidities: hypertension 59%, diabetes 8%, coronary heart di
Read more

HDL a negative risk factor? or cholesterol efflux??

Although the data on HDL being protective of atherosclerotic disease is pretty consistent, there are negative studies, and the data on medications which dramatically improve HDL levels (eg, the cholesteryl ester transfer protein --CETP --inhibitors, such as torcetrapib) are clinically disappointing. there have been several explanations for this. for example, some HDL particles are "pro-inflammatory" and elicit an atherosclerotic response (which may be related to apolipoprotein C-III). a new article in NEJM highlights another angle -- that the issue is not the HDL number, but HDL's functionality in the reverse transport of cholesterol, as measured by the cholesterol efflux capacity (see  lipids HDL efflux capacity vs amt NEJM 2014 in dropbox, or  DOI: 10.1056/NEJMoa1409065). This article looks at the Dallas Heart Study, where they measured HDL levels, HDL particle concentrations, and cholesterol efflux capacity in 2924 adults free of cardiovascular disease and followed p
Read more

UPDATE: ASCVD risk factor critique

I decided to do an extensive update of my prior ASCVD (atherosclerotic cardiovascular disease) risk analysis and critique of risk calculators, with much more information about the “non-traditional” risk factors. I felt that there should be more data/clarification, since I think we are really undertreating/not preventing the most common cause of death by relying on the current ASCVD risk calculators   -- Cardiovascular risk models:      -- many of us rely on the various cardiovascular risk models to determine the appropriate approach to patients, and these are often integrated into the electronic medical records      -- these risk models have several very important limitations:          -- they are typically based on community data, eg from the Framingham study etc. But, in clinical practice, we are dealing with the individual and not the community:              -- there is very clear evidence that there are many individual risk factors that are associated with increased risk of a cardi
Read more