doxycycline has fewer C diff infections than azithro

 One concern with many clinical studies is that their focus is too focused: eg, does this antibiotic cover a certain infection, and is it better than other antibiotics? However, there are broader concerns in general for drugs. Do they create the basis for other problems, perhaps even worse than the initial condition? A recent VA study found that doxycycline resulted in significantly fewer cases of Clostridioides difficile infection (AKA, Clostridium dificile of old, or just CDI) vs azithromycin: see cdiff doxy better than azithro AmJInfControl2023 in dropbox, or  https://doi.org/10.1016/j.ajic.2023.09.007

 

Details:

-- 156,107 hospitalized patients in the VA system were diagnosed with community-acquired pneumonia (CAP) in the large VA database, for this retrospective study of patients from 2009-2022

        -- 87% (135,703 patients) received azithromycin and 13% (20,404 patients) doxycycline, both along with ceftriaxone

-- 96% male

-- comparing those on doxycycline vs azithromycin, there were several differences: age 72.8yo vs 71.1yo (78% >65yo vs 71.2%), Charlston score for comorbidities 5.4 vs 4.8 (45.5% having high score of 6-23 vs 38.3%), prior CDI by PCR in past year 1.1% vs 0.6%

-- comorbidities: MI 22.6% vs 17.4%, heart failure 49.4% vs 40.4%, diabetes without complications (31.7% vs 26.9%), diabetes with complications (49.2% vs 45.4%), dementia (14.2% vs 10.2%), COPD (69.9% vs 68.6%), PUD (9% vs 8%), mild liver disease(17.4% vs 15.2%), severe liver disease (2.9% vs 2.3%), renal disease (33.7% vs 28.3%), cancer (34.5% vs 33.3%)

    -- ie: those on doxycycline were older and all comorbidities were worse: all comparisons favored the azithromycin group, with p<0.0001

-- PPI use: 45.4% of those on doxycycline vs 41.7% on azithromycin, p<0.0001 (more PPI use also puts the doxycycline group at higher risk of C. diff infection: see below)

--there was no difference between the groups in the number of doses of ceftriaxone

 

-- Main outcome: the incidence of C. difficile infection (CDI) in patients with community-acquired pneumonia (CAP) treated with ceftriaxone plus doxycycline versus azithromycin

 

Results:

 -- 1234 patients developed CDI within 30 days of starting antibiotics

    -- the overall rates of CDI: 1085 patients on azithromycin vs 149 patients on doxycycline 

    -- doxycycline vs azithromycin: 17% decreased odds ratio of developing a CDI within 30 days of pneumonia treatment, OR 0.83 (0.70-0.99), p=0.0348

 

-- subgroup analyses:

    -- for all patients with  CDI in the past year: chance of new CDI 10.5% vs 0.6%, p<0.0001, 18.5 times the risk of developing CDI in adjusted multivariable logistic regression model

    -- comparing patients on doxycycline vs azithromycin:

        -- patients  with prior CDI (1,064 patients): in doxycycline group 1.1%, in azithromycin group 0.6%

            -- 45% decrease with doxycycline, OR 0.55 (0.31-0.91), p=0.0179

        -- patients without prior CDI (155,043 patients): nonsignificant trend of decrease with doxycycline, OR 0.88 (0.73-1.05)

-- overall,  “doxycycline was the only factor associated with decreased CDI incidence”

 

    -- for all patients, those with a higher Charlson Comorbidity Index: 20% increased risk of CDI, OR 1.2 (1.08-1.36)

         -- Charlson Comorbidity Index of 6-23: 22% risk of CDI, OR 1.22 (1.1-1.4), p=0.0011

            -- ie, the study confirmed the role of more comorbidities being associated with a higher CDI risk

 

    -- no difference in CDI by the age of the patient being >65 vs <65yo

 

-- Commentary:

-- CDI is associated with 15,000 to  30,000 annual deaths per year in the US, and about 35% of patients who develop CDI have recurrent episodes (which can be progressively  harder to treat)

-- the Infectious Disease Society of America guidelines recommends treating CAP with a beta-lactam antibiotic along with a macrolide, with doxycycline as “an alternative” to the macrolide

 

-- this study did find that doxycycline was associated with a 17% decreased risk of CDI vs azithromycin, both given with ceftriaxone for community-acquired pneumonia.

    -- this was despite the fact that the patients on doxycycline were twice as likely to have had a history of CDI in the prior year, were almost 2 years older, and had significantly higher rates of all of the measured comorbidities (all of which predispose people to a subsequent CDI infection); and those on doxycycline also had more PPIs prescribed, which is also associated with CDI in several studies (https://gmodestmedblogs.blogspot.com/2017/04/ppis-and-recurrent-c-diff-infections.html)

 

-- other studies have found similar results:

    -- in a five-year period, 2305 patients at San Francisco General Hospital who had received ceftriaxone for different conditions including pneumonia were evaluated for the incidence of CDI comparing those who also received doxycycline versus only ceftriaxone, finding that the incidence of CDI was 1.67 cases per 10,000 patient days in those who received the additional doxycycline versus 8.11 per 10,000 patient days in those patients who did not, with an adjusted multivariable model finding that for each day of doxycycline received, the rate of CDI was 27% lower in those who also received doxycycline (see Clin Infect Dis. 2012 Sep;55(5):615-20 doi: 10.1093/cid/cis457)

    -- another case controlled study found that in patients who had hospital-acquired CDI from 1999 to 2005, where all patients had received antibiotics in the prior 60 days, the odds ratio of acquiring CDI decreased progressively by antibiotics given, from imipenem-cilastin (OR 2.77), to clindamycin (OR 2.31), moxifloxacin (OR 1.88), and ceftriaxone (OR 1.49).  However, metronidazole was associated with a decreased risk (OR 0.67), and doxycycline with the lowest risk (OR 0.41). The study also found that the usual suspects for CDI (use of proton pump inhibitors, longer hospital stays, older age, prior gastrointestinal disease, and prior non-CDI infection) were associated with increased risk of CDI (https://pubmed.ncbi.nlm.nih.gov/18171186/ )

 

-- the big issue here (to me) is the non-specificity of even effective antibiotic therapy: it can treat the acute infection (CAP, in this case) but can also have dramatic effects on the gut microbiome. And a healthy gut microbiome is so important for so many known (and undoubtedly still unknown) healthy physiological functions in the human body, as examples:

    -- proton-pump inhibitors, as a non-antibiotic example, are associated with major gut microbiome changes in several studies, presumably related to the depletion of helpful bacteria as well as undeterred growth of harmful oral bacteria that might otherwise have had limited access to the intestines if there were normal acidity in the stomach: https://bmcmicrobiol.biomedcentral.com/articles/10.1186/s12866-023-02895-w#:~:text=The%20effects%20of%20PPIs%20on,is%20mediated%20by%20the%20increase

        -- and, as noted above, PPIs are associated with an increased risk of CDI: https://gmodestmedblogs.blogspot.com/2017/04/ppis-and-recurrent-c-diff-infections.html

    -- there are some interesting studies finding that metformin augments a few commensal bacterial species in the gut microbiome, especially Akkermansia (https://www.nature.com/articles/s41540-022-00261-6#:~:text=Metformin%20alters%20the%20gut%20microbiome,butyrate%20and%20propionate7%2C8), and that this bacterium actually improves glucose tolerance and may well explain at least part of the benefit of metformin. An RCT of metformin vs placebo for 4 months in patients with type 2 diabetes found a rapid initial change in the microbiome in those on metformin that paralleled their decrease in A1c levels, and on transplanting these human fecal samples into germ-free mice, glucose tolerance improved in the mice but only in those receiving fecal samples from the group of people on metformin (https://www.nature.com/articles/nm.4345). A prior mouse study found that oral administration of Akkermansia in the absence of metformin in  germ-free mice led to improved glucose tolerance (see dm metformin microbiome GUT 2014 in dropbox, or doi.org/10.1136/gutjnl-2012-303839) .

     -- the gut/brain axis (see microbiome gut brain axis JClinInvest2015 in dropbox, or doi:10.1172/JCI76304) has been pretty well described showing that the gut microbiome can affect emotional behavior, stress- and pain-modulation, and brain neurotransmitters (the gut microbiome elaborates many neurotransmitters); animal studies have found major effects in the gut microbiome on the expression of brain signaling systems

     -- as another example, consuming L-carnitine (as found in red meats, and chicken to a much lower extent) in omnivores led to very high systemic levels of the profound atherogen TMAO; but in vegans and vegetarians given L-carnitine, there was much less TMAO produced. Further research found that there were specific bacteria in the gut that were associated with the TMAO elevations, suggesting a 2-step process: non-meat eaters when eating meat (or L-carnitine) develop a selection of specific bacterial taxa that on subsequent red meat consumption leads to increased TMAO: see cad red meat TMAO nature medicine 2013 in dropbox, or doi:10.1038/nm.3145 as referenced in https://gmodestmedblogs.blogspot.com/2023/10/red-meat-increases-risk-for-diabetes.html

    -- it is important to realize that the relationship between antibiotics and the gut microbiota may well be bidirectional: the gut microbiota may be radically altered by the destructive effect of the antibiotics on beneficial microbes and increased growth of opportunistic pathogens; on the other hand these microbiome changes may influence the absorption and metabolism of the antibiotics and their efficacy

-- one important point here is that microbiome is dynamic, it changes depending on the situation at hand: unhealthy behaviors (eg sedentary lifestyle, unhealthy eating, stress, anxiety, depression, destructive antibiotics) hurt it, and healthy behaviors improve it, though may take months to recover. For example, diet (especially lots of green veges, less sugar/processed foods), exercise, pre- and pro-biotics (the former includes lots of the better foods listed, as well as beans, lentils, chickpeas, garlic, bananas, cocoa, apples, whole grains, apples, onions) all lead to an improvement in the health of the microbiome

    -- probiotics given with antibiotics may also help improve gut microbial diversity and composition (https://www.microbiologyresearch.org/content/journal/jmm/10.1099/jmm.0.001625 ) and, though the quality of the studies and the lack of standardization of probiotics do raise real questions, probiotics may well decrease both antibiotiic-associated diarrhea and CDI (https://jamanetwork.com/journals/jama/fullarticle/1151505)

 

Limitations:

-- this was a retrospective massive data-mining study, raising a few concerns:

    -- it is not clear why the azithromycin vs doxycycline was chosen by the clinicians for the pneumonia therapy. It was pretty striking that those on doxycycline were at much higher CDI risk given their comorbidities as well as being prescribed more PPIs. But, why were thee individuals more likely to be given doxycycline??

        -- and, perhaps most importantly, those on doxycycline were twice as likely to have had a history of CDI in the past year (1.1% vs 0.6% on azithromycin, p<0.0001). 

        -- ie, this all seems a far stretch from being random antibiotic allocation, which raises the large question of selection bias (though in this case, these biases themselves would predict much more CDI with doxycycline, yet azithromycin was much worse for developing CDI: ie, a balanced, randomized study would likely have shown doxycycline to be even much better than azithro than what was found in this study…..)

    -- any conclusions need to be considered associations and not causality, since this was an observational study given the potential for unmeasured confounders

        -- for example, there was no info on meds taken, smoking, exercise, diet, weight/BMI, social circumstances (crowded living conditions, etc), stress, other psychological/psychiatric diagnoses.  And all of these items could affect the microbiome

    -- the comorbidities they assessed were all binary: no information on the seriousness of the heart failure, diabetes, renal disease…..  And these gradations are important in considering the adverse effects of the comorbidities

-- this was a population of male veterans, and the results may not be generalizable to the overall population

-- no information on baseline C. difficile carriage: other studies have found that 20% of hospitalized patients are asymptomatic C diff carriers; up to 50% in those in long-term care facilities (eg, see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4647607/ , or https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4133563/ with no difference in those with recent hospital admission, or antimicrobial exposure within 90 days). Though the % of patients developing clinical CDI in this VA study was low (<1%), generalizing these results to other communities with higher baseline C difficile carriage rates may be fraught

    -- a study in Denmark found that community-based individuals with "unformed stools" had a 1.5% rate of CDI and only 61% of them had prior antibiotic use or hospitalization: https://gmodestmedblogs.blogspot.com/2014/08/community-c-diff-infections.html

        -- given the overlap in this Danish study with people having IBS (irritable bowel syndrome), it is interesting that probiotics may well help people with IBS: https://gmodestmedblogs.blogspot.com/2014/12/probiotics-in-irritable-bowel-syndrome.html

 

So, this VA study does suggest a few things:

-- it is very important to look beyond the efficacy of antibiotics in treating infections, and that the longer-term consequences (eg changes in microbiome) are really important to assess and be included in our decision tree for antibiotic selection

    -- which means that a somewhat less potent antibiotic may, in an overall assessment, be a better one for patients

-- the microbiome is remarkably important in promoting and preserving a healthy human physiology

    -- and, the microbiome, fortunately, is a dynamic entity, with important beneficial changes induced by a healthy diet, exercise, and avoiding things that injure it. Though it may take several months to recover

 

The next blog will review a systematic review/meta-analysis finding that doxycycline had comparable effects in patients with mild-to-moderate CAP vs macrolide antibiotics

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:

 

 go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- 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

 

if you would like to see the articles in this blog, please email me. 

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list