Antibiotic use in kids: narrow-spectrum better than broad-spectrum

a retrospective study of children treated for acute respiratory tract infection with antibiotics found that broad-spectrum ones did no better in treating the infection and had more adverse events (see doi:10.1001/jama.2017.18715​).

Details:
--30,159 children aged 6 mo to 12 yrs were prescribed antibiotics for an acute respiratory tract infection in 2015, in a retrospective cohort study from a network of 31 pediatric primary care practices with 336 clinicians in New Jersey and Pennsylvania
--additionally, 2472 kids were followed prospectively between January 2015 to April 2016
--children were stratified by propensity score matching, to mathematically equalize patient-level characteristics so that the kids were pretty similar in the different groups
--qualitative interviews were conducted on 109 legal guardians and 24 children at 4 practices, assessing child suffering, missed school or work, sleep quality, adverse events, speed of symptom resolution, using the Pediatric Quality of Life Inventory
--adverse events and treatment failure were assessed through 14 days (primary outcome) and 30 days after diagnosis. treatment failure was defined as a kid getting a new systemic antibiotic prescribed for the same acute respiratory tract infection; though if a new antibiotic were prescribed within the time-frame of the treatment for the initial infection, the records were reviewed to see if that were due to an adverse event or treatment failure

Results:
--retrospective cohort:
    --19179 kids were treated for acute otitis media; 6746 for group A streptococcal pharyngitis, and 4234 for acute sinusitis
    --4307 (14%) were prescribed broad-spectrum antibiotics (amoxacillin-clavulanate, cephalosporins, macrolides) vs the rest prescribed narrow-spectrum ones (penicillin or amoxicillin)
    ​--broad spectrum antibiotics were not superior to narrow-spectrum ones; though on results stratified by diagnosis, there was a small but significant decrease in treatment failure in those on broad-spectrum antibiotics (1% vs 2.4%) but only for Group A strep infections
    --adverse events through 14 days: 3.7% for broad-spectrum and 2.7% for narrow-spectrum antibiotics (p<0.001 for fully matched analysis), but not significant at 30 days (8.7% vs 8.1%)
    --treatment failure through 14 days: 3.4% vs 3.1% for broad vs narrow-spectrum, nonsignificant; and 8.7 vs 8.1% at 30 days, also nonsignificant
--prospective cohort:
    --1100 were treated for acute otitis media; 705 for group A streptococcal pharyngitis, and 667 for acute sinusitis
    --868 (35%) were prescribed broad-spectrum antibiotics
    --broad-spectrum antibiotics were associated with slightly worse child quality of life (score 90.2 vs 91.5 for narrow-spectrum, which was statistically significant, but probably not clinically so, since 4 point differences are considered to be meaningful), but no difference in other patient-centered outcomes
--adverse events: clinician-documented 3.7% for broad-spectrum vs 2.7% narrow-spectrum; patient-reported 35.6% for broad-spectrum vs 25.1% for narrow-spectrum; of those guardians or patients reporting adverse events, 70% were from diarrhea, 40% rash, 21% nausea and/or vomiting, 28% with more than 1 adverse event

Commentary:
--antibiotics are the most commonly prescribed med for kids and the majority of antibiotics are for acute respiratory infections
--currently about 1/2 of these prescriptions are for broad-spectrum antibiotics
--the American Acad of Pediatrics does recommend penicillin or amoxicillin as first-line therapy for otitis media and amoxicillin for acute sinusitis (as per the 2015 Red Book), though some other societies promote more broad-spectrum antibiotics. And, newer clinical trials tend to use the broad-spectrum antibiotics, implicitly suggesting to clinicians that these are the best drugs
--although propensity score matching does help equalize the kids' different characteristics, they did not address confounding by clinician characteristics (and there are data supporting the conclusion that physicians can have very different antibiotic prescribing styles, independent of the apparent clinical presentation of the child)
--the only positive clinical difference for broad-spectrum antibiotics in the above study was for strep pharyngitis, but this was NOT FOUND in the prospective cohort, and the difference in the retrospective one was really not clinically important (99% vs 97% effective in this subgroup). and, though not significant, the trend for the other respiratory illnesses was in favor of the narrow-spectrum ones...
--also, an interesting finding (not so surprising) was that clinician-reported adverse events was about 1/10th that of the guardian-reported ones: ie, be wary of the adverse events reported in many studies; many may not even be reported (or documented) by the clinician. And, this might distort our views as clinicians in that it may frequently be the case that a drug has an adverse effect, we do not hear about it, so we blithely/unwittingly go one prescribing the drug thinking all is fine.....
--the drug company argument was that non-typical bacterial strains associated with upper respiratory infection (eg Moraxella catarrhalis, H influenzae and Staph aureus) should have antibiotic coverage as well. this is pretty clearly undercut by the above study (it was never really clear to me, at least, that several of these organisms were actually causative for infection, vs just being fellow-travelers/hangers-on)
--though an evident limitation of this study is that there was usually no documentation of the infectious cause of these respiratory infections, the results do reflect clinical practice overall since we treat patients basically by their symptoms and signs and not their bacteriology (which we rarely know, except with strep pharyngitis)

so, impressive study in that it confirms that drug companies are indeed effective in swaying clinicians!!! i remember reading reports many years ago specifically about how the intensive drug rep blitz promoting the use of azithro for acute respiratory infections led to rapid and dramatic increases in its use. now those chickens (and not antibiotic-free ones) are coming home to roost: we now have more people getting less good antibiotic treatment with more short-term adverse effects reported, with more antibiotic resistance (see blogs below) and more dramatic microbiome changes (some of which may persist well after the antibiotics are gone).  and, this brings up tangentially another of my drug company concerns: they are the ones designing, carrying out, and publishing increasing percentages of articles reported in the premier medical journals, and, as per many prior blogs, many of the study details are not transparent and may bias the results in the direction the drug company wants (eg, studies where control patients may be getting inferior drugs to compensate for the positive effects of the study drugs, thereby exaggerating the stated benefit of the drug company's new drug (eg: see http://gmodestmedblogs.blogspot.com/2015/12/empagliflozin-good-and-bad.html )

for other relevant blogs:
see http://gmodestmedblogs.blogspot.com/2015/07/antibiotic-overprescribing.html  for blog on similar issue of antibiotic overuse, showing the trend to decreasing use of narrow-spectrum antibiotics and increase in macrolides, as well as a CDC report showing that there was increasing group A strep resistance to macrolides (10%): ie, using an unnecessarily broad-spectrum antibiotic with increasing likelihood that it will be ineffective!!!

see http://gmodestmedblogs.blogspot.com/2015/11/longterm-microbiome-changes-with.html  for a blog finding prolonged microbiome changes (12 month) after a single exposure to antibiotics

see http://gmodestmedblogs.blogspot.com/2017/04/antibiotics-microbiome-changes-and.html  ​ for one demonstrating increase in colorectal adenomas, including advance ones, in women exposed to antibiotics

and, of course, the array of microbiome blogs: see http://gmodestmedblogs.blogspot.com/search?q=microbiome​