Without evidence of benefit, an intervention should not be presumed to be beneficial or safe.

- Rogue Medic

More evidence that antibiotics are over-prescribed – How should that change?

ResearchBlogging.org
 

Today Lancet Infectious Diseases posted an early release of an article that shows that the antibiotic amoxicillin still does not work on viruses.

Many doctors still routinely prescribe antibiotics for viral infections.

Bias, perhaps the biggest confounder in medical research, is controlled for very well.

 

Acute uncomplicated lower-respiratory-tract infection is the most common acute illness managed in primary care in developed countries; most patients receive antibiotics, even in low-antibiotic-prescribing countries.1, 2, 3 and 4 [1]

 

Will this be used by alternative medicine pushers as if it is evidence that alternative medicine works?

Yes.

Is this evidence that alternative medicine works.

Absolutely not.

This is only evidence that medicine needs a lot of improvement, not that medicine needs to be replaced by quackery.
 

Eligible patients were aged 18 years or older and consulting for the first time with either an acute cough (≤28 days’ duration) as their main symptom, for which non-infective diagnoses were judged very unlikely, or an illness in which cough was not the most prominent symptom but the clinician thought acute lower-respiratory-tract infection the most probably diagnosis.[1]

 

What exclusions might have affected the outcome?
 

We excluded patients in whom the initial clinical diagnosis was community-acquired pneumonia17 (ie, complicated lower-respiratory-tract infection) on the basis of focal chest signs (focal crepitations, bronchial breathing) and systemic features (high fever, vomiting, severe diarrhoea). We did not use a formal clinical prediction rule for the diagnosis of pneumonia because of the absence of consensus about which rule to use.18, 19, 20 and 21 [1]

 

Is it wrong to exclude a disease that is known to benefit from the treatment?

No. This was not a study of whether amoxicillin is efficacious for pneumonia. This was a study of whether amoxicillin is efficacious for acute lower respiratory tract infections that are not pneumonia, or that are not suspected to be pneumonia. There is no reason to use any drug other than amoxicillin for this condition.[2]

Patients were also excluded if there was no suspicion that the cause was due to an infection (pulmonary edema, reflux, allergy, et cetera).

 

Previous diagnoses of asthma, chronic obstructive pulmonary disease, and other comorbid disorders were not exclusion criteria, and thus acute infective exacerbations were included.[1]

 

The most important part –
 

Both clinicians and patients were masked to the randomisation sequence, and all outcome data were gathered masked to allocation status.[1]

 

Only the manufacturer knew which was placebo and which was amoxicillin. There was emergency access to the information only in cases of medical emergencies. They do not state how many times this information was given out, but they later state that only three patients overall were acutely admitted to the hospital.
 

The amoxicillin and placebo were manufactured to be identical in appearance, taste, and texture.[1]

 

Those who agreed to randomisation were allocated (1:1) to receive three times daily either 1 g amoxicillin or placebo for 7 days by the clinician, . . .

We estimated that, to achieve bacterial eradication, concentrations needed to be higher than the minimum inhibitory concentration for at least 5 days. We chose a 7 day course for acceptability to clinicians and to allow for poor compliance.[1]

 

A diary of symptoms and severity was maintained for up to 28 days.
 

Members of the research team telephoned participants after 4 days to offer support and answer any questions about completion of the diary. If the diary was not returned after 4 weeks, we collected brief information with either a short questionnaire or a standardised telephone call about symptom duration and severity. Participating clinicians registered all contacts with patients for 4 weeks after the initial consultation, including referral to hospitals and out-of-hours contacts.[1]

 

We used Stata (version 11.2) for all analyses. For interpretation we used a significance threshold of 5% for the primary outcome and 1% for the secondary outcomes to minimise the chance of type I error with multiple outcomes.[1]

 

They did add analysis of patients over 70 after the completion of data collection, so that should not be seen as a part of the prospective study. They do seem to understand the importance of minimizing the possibility of having a lot of secondary endpoints produce statistically significant, but therapeutically meaningless results.
 

The funding sources had no roles in data collection, analysis, or interpretation; report writing; or submission.[1]

 

Imagine if there were that kind of independence with all research.
 

 
Click on images to make them larger.

Not many patients were lost to followup for a study of this kind.
 

 

The patients appear to have been randomized well.
 

 

Interpretation

Our trial is the largest study so far of the use of antibiotics in acute lower-respiratory-tract infection, and adds much to the placebo-controlled evidence noted in the Cochrane review, especially data for patients aged 60 years or older.16 Compared with placebo, amoxicillin did not significantly affect the duration of symptoms rated “moderately bad” or worse in the first few days of infection, neither overall nor in patients older than 60 years. Symptom severity also did not differ significantly between treatment groups. Amoxicillin prevented some patients from developing new or worse symptoms but the number needed to treat was high and matched by a similarly sized number needed to harm for side-effects. Our data suggest, if anything, a smaller benefit from antibiotics for symptoms and a clearer estimate of harms than did the Cochrane review. Thus, unless pneumonia is suspected, antibiotics should not be prescribed for patients with acute lower-respiratory-tract infection.[1]

 

When the NNT (Number Needed to Treat) and the NNH (Number Needed to Harm) are the same, we are not providing a benefit to patients.

We are just changing the symptoms.

In this study, the differences are subtle and not statistically significant.

When the difference is so small that it does not reach statistical significance in a large study, it is unlikely that it will be significant to patients.[3]
 

In the absence of evidence of benefit, there is no good reason to be using a treatment outside of a controlled trial.

Footnotes:

[1] Amoxicillin for acute lower-respiratory-tract infection in primary care when pneumonia is not suspected: a 12-country, randomised, placebo-controlled trial.
Paul Little et al.
The Lancet Infectious Diseases, 2012
DOI: 10.1016/S1473-3099(12)70300-6

[2] Treatment of acute cough/lower respiratory tract infection by antibiotic class and associated outcomes: a 13 European country observational study in primary care.
Butler CC, Hood K, Kelly MJ, Goossens H, Verheij T, Little P, Melbye H, Torres A, Mölstad S, Godycki-Cwirko M, Almirall J, Blasi F, Schaberg T, Edwards P, Rautakorpi UM, Hupkova H, Wood J, Nuttall J, Coenen S.
J Antimicrob Chemother. 2010 Nov;65(11):2472-8. doi: 10.1093/jac/dkq336. Epub 2010 Sep 18.
PMID: 20852271 [PubMed – indexed for MEDLINE]

Free Full Text from Journal of Antimicrobial Chemotherapy

RESULTS:
Compared with amoxicillin, no antibiotic class (and no antibiotic treatment) was associated with clinically relevant improved symptom resolution (all coefficients in the range -0.02 to 0.01 and all P values greater than 0.12). No antibiotic class (and no antibiotic treatment) was associated with faster time to recovery than amoxicillin.

 

 

While the differences with co-amoxiclav (Augmentin) and sulphonamides/trimethoprim (Bactrim) may seem good, they are based on very small numbers. Cephalosporins also give the appearance of doing better than amoxicillin, but nothing reached statistical significance. When the difference is so small that it does not reach statistical significance in a large study, it is unlikely that it will be significant to patients.

[3] Fiction And Fantasy In Medical Research: The large scale randomised trial
By James Penston
Published by London Press, 2003
ISBN 0954463617, 9780954463618
144 pages
GoogleBooks link

Little, P., Stuart, B., Moore, M., Coenen, S., Butler, C., Godycki-Cwirko, M., Mierzecki, A., Chlabicz, S., Torres, A., Almirall, J., Davies, M., Schaberg, T., Mölstad, S., Blasi, F., De Sutter, A., Kersnik, J., Hupkova, H., Touboul, P., Hood, K., Mullee, M., O’Reilly, G., Brugman, C., Goossens, H., & Verheij, T. (2012). Amoxicillin for acute lower-respiratory-tract infection in primary care when pneumonia is not suspected: a 12-country, randomised, placebo-controlled trial The Lancet Infectious Diseases DOI: 10.1016/S1473-3099(12)70300-6

Butler CC, Hood K, Kelly MJ, Goossens H, Verheij T, Little P, Melbye H, Torres A, Mölstad S, Godycki-Cwirko M, Almirall J, Blasi F, Schaberg T, Edwards P, Rautakorpi UM, Hupkova H, Wood J, Nuttall J, & Coenen S (2010). Treatment of acute cough/lower respiratory tract infection by antibiotic class and associated outcomes: a 13 European country observational study in primary care. The Journal of antimicrobial chemotherapy, 65 (11), 2472-8 PMID: 20852271

.