The Futile Quest of Proofing a Negative in the Presence of Overwhelming Absence of Harm

The cartilage damage caused by fluoroquinolones in juvenile beagle dogs is not subtle:  blister formation and erosions occur in a dose-dependent fashion resulting in reversible cartilage damage.  beagle dogThese lesions can be seen not just histologically but even macroscopically.  The animals are unable to walk as weight-bearing joints are predominantly affected.   For levofloxacin and ciprofloxacin, the threshold or NoEL was found to be 3 mg/kg/day and 10 mg/kg/day, respectively, when the drug was administered for 14 days; below this dose joint damage did not occur in the beagle dog model [1, 2],.  We are not aware of any other animal species that has a similar toxic response.

Now, Bradley and colleagues from J&J just published the results of a prospective trial in which children were followed for 5 years after the initial exposure to a fluoroquinolone (FQ), levofloxacin [2].  This study was an extension of a larger 1-year safety trial; it concentrated on a subset of ‘higher-risk’ patients identified by a list of criteria which included growth retardation, abnormal bone or joint symptoms/signs, or persistent musculoskeletal pain.

Studies of this nature are difficult to execute, and this one was no exception.  Compliance is an issue because of the long-term follow-up period, and there always is concern about bias.  In this case, neither parents nor investigators were blinded, but members of the DSMB of the 5-yr extension cohort supposedly were.  Establishing whether there existed a relationship between an AE and the study drug several years after initial exposure also seems highly problematic.  One could call it ‘expert guesswork’.

The results came out as expected and were anticlimactic.  Of the enrolled 124 children with previous exposure to levofloxacin and the 83 comparator subjects only half completed the 5-yr observation period in each arm.  There were 3 cases of musculoskeletal AEs in each group.  In the levofloxacin group, hypermotility (sic!), arthralgia, and arthritis were noted, mostly mild in nature.  However, 1 patient developed severe synovitis.  There was no difference in growth between the 2 groups.  Taken together, there was no safety signal of concern.

The results of this trial will sway nobody’s mind regarding the role of FQ in pediatrics.  Even if there were a very small risk of arthropathy, the incidence must be exceedingly small making it impossible to rule it out or estimate its frequency with any precision.  A prospective controlled clinical trial is a very blunt tool and unsuitable to address the question at hand.

After 30 years of FQ use, and hundreds of million doses prescribed, one would assume that even rare AEs would eventually have declared themselves and become manifest.  Such was the case for azithromycin (cardiac side effects) and some other drugs like tramadol (also cardiac events).

As the use of antibiotics for AOM is now much more restricted, concerns about a possible ‘abuse’ of FQ for trivial viral infections should be much less of an issue nowadays [3].  With minimal harm, if at all, and in light of more appropriate use of antibiotics in children which now includes ‘watchful waiting’, and no evidence of persistent arthropathy, the pediatric labeling in the Package Insert should be revised.

Maybe, just maybe, a fresh look at the risk-to-benefit ratio would result in wording that states that prolonged or prophylactic use is not appropriate.  In its current form, the FQ Warnings/Precautions on pediatric use seem overly restrictive and unbalanced.


[1] Von Keutz.  Arch Toxicol. 2004; 78(7):418
[2] J Bradley.  Pediatrics 2014;134:e146
[3] A. Lieberthal.  Pediatrics 131, 2013 pp. e964

Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published. Required fields are marked *