Oral antibiotics are noninferior to IV in pediatric bone and joint infections (obviously)

Some topics come up over and over again, and it may seem somewhat repetitive or wasteful to spend so much time on them, but seeing as so many people are still using outpatient IV antibiotics despite overwhelming evidence that oral antibiotics are just as good, if not better, I will continue to cover papers as they arrive. That brings us to a large multicentre RCT of pediatric bone and joint infections. 

The paper

Nielsen AB, Holm M, Lindhard MS, et al. Oral versus intravenous empirical antibiotics in children and adolescents with uncomplicated bone and joint infections: a nationwide, randomised, controlled, non-inferiority trial in Denmark. Lancet Child Adolesc Health. 2024 Jul 15:S2352-4642(24)00133-0. doi: 10.1016/S2352-4642(24)00133-0. PMID: 39025092 NCT04563325

The Methods

This is a multicentre, randomized, controlled, open-label, non-inferiority trial from Denmark.

Patients

Pediatric patients from 3 months to 17 years in age with a suspected bone and joint infection. “Criteria for suspected BJI included local signs of infection (eg, pain, impaired function, warmth, redness, or swelling) in combination with fever or elevated inflammatory markers and the absence of other apparent differential diagnoses.” They enrolled patients early, based only on suspicion of infection, because acute therapy is needed, but removed patients from the trial if an alternative non-infectious diagnosis was later made. 

They excluded patients with “severe or complex” infections, such as septic shock, “substantial soft tissue infection”, prosthetic material, patients with a previously documented infection with a resistant pathogen with few oral options (such as MRSA), or the need for surgery within 24 hours. 

Intervention

Oral high-dose antibiotics

• For children under 5 years: coformulated (8:1) amoxicillin (100 mg/kg per day) and clavulanic acid (12·5 mg/kg per day) in three doses.
• For children 5 and older: dicloxacillin (200 mg/kg per day, up to a maximum of 8000 mg per day) in four doses.

Comparison

Intravenous antibiotics: ceftriaxone (100 mg/kg per day, up to a maximum of 4000 mg per day) in one dose.

Shared procedures

In both groups, the treating clinician could change to a different antibiotic using the same route if warranted, such as allergic reaction, bad taste, or culture results. 

If staph aureus was confirmed, oral rifampicin (20 mg/kg per day, up to a maximum of 900 mg/day) in three doses was allowed to be added in both groups.

After a minimum of 3 days, and with signs of clinical improvement, both groups were switched to “standard dose oral antibiotics”, which was just a lower dose of the Clavulin or dicloxacillin. 

Total treatment length was a minimum of 1 week for joint infections, 3 weeks for bone infections, and 4 weeks for spondylodiscitis or sacroiliitis.

Outcome

The primary outcome was clinical sequelae at 6 months, defined as atypical mobility or function of the affected bone or joint.

They started their trial with a non-inferiority margin of 10%, meaning they were willing to accept oral antibiotics if they were up to 10% worse than IV. When none of the first 100 patients had any sequela, they realized this margin was way too big, so they adjusted their non-inferiority margin down to 5%. 

The Results

They enrolled 248 patients, 54 of whom did not actually have a bone or joint infection and so were subsequently removed from the trial. These are mostly younger kids, with 67% falling into the 0-4 year age range. 90% of the bone infections were confirmed by MRI. Joint infections were mostly “confirmed” by MRI or ultrasound, although I would have thought joint aspiration was the real gold standard. 

IV antibiotics were given for a median of 3.3 days, while the higher dose oral antibiotics were given for a median of 4.3 days. 

17% of the oral group was switched to IV. 9 patients switched because they weren’t tolerating oral antibiotics, 3 because they had surgery, 1 because of a resistant pathogen, and 4 because of suspected treatment failure. 9% of the intravenous group switch to oral because of difficulties getting an IV line. 4 children (2%) had IV antibiotics for the entire duration of therapy because they couldn’t tolerate oral at all.

For the primary outcome, clinical sequelae at 6 months, there were 0 outcomes in either group. I don’t know if it is really fair to run statistics with 0 outcomes, but the numbers they run show a one sided 97.5% confidence interval up to 3.8%, meaning that they can conclude that oral antibiotics are noninferior to IV. Of course, with 0 outcomes in both groups, it is just as likely that IV antibiotics are worse than oral, making this one sided test silly. 

For their primary safety outcome, no child in either group had a need for admission to an intensive care unit, septic shock, organ failure, or endocarditis.

They talk about some sensitivity analyses, such as looking only at patients who didn’t cross-over, but with 0 outcomes in either group, those analyses are obviously silly. 

There are some differences in adverse events, but as I discuss below, the adverse events they decided to look at are tremendously biased. (Nowhere in this manuscript do I see any mention of IV complications or the total number of times these children were poked with a needle.)

My thoughts

This is an important trial that is unfortunately undone by a poorly chosen outcome. I don’t manage these infections long term, so I have no idea where they went wrong. Did they just get incredibly unlucky to have 0 patients with long term clinical sequelae? When they started this trial, they expected 5% of patients to have bad outcomes. Was that based on clinical experience, a literature review, or were they just randomly guessing? Unfortunately, whoever was designing this trial woefully misunderstood the clinical course of bone and joint infections in this pediatric population, and therefore chose an inappropriate primary outcome that we can’t really use to make clinical decisions.

It is definitely nice to know that these kids will all be fine in 6 months no matter what we do. It really makes you question their exclusions, and the patients who crossed over from oral to IV antibiotics. These outcomes are close to miraculous. With no patients having any clinical sequelae at 6 months, these infections seem no more dangerous than the common cold. (Hell, COVID has more than 0% 6 month sequelae.) 

Their exclusion of patients with “substantial soft tissue infection” was almost certainly unnecessary, based on the available literature – The OVIVA trial only looked at complex orthopedic infections, and showed that oral was non-inferior to IV – and unfortunately muddies the waters here. It adds unnecessary subjectivity to the results. What exactly is a substantial infection? On the leg of a 3 month old, 4 centimeters of cellulitis takes up a huge percentage of the leg, and I guess could seem substantial, but clearly is treated with oral antibiotics. Fans of intravenous antibiotics (ie, those untutored in science), will use this exclusion as a get out of jail free card in justifying their use of IV antibiotics.

I am also not sure it makes sense to exclude patients with MRSA, because that is a pathogen we all see frequently, and mostly can be treated with oral antibiotics such as trimethoprim-sulfamethoxazole. However, it does make the data somewhat cleaner, so I understand the decision. 

I don’t really understand using 2 different doses of oral antibiotics. If the higher dose is effective and tolerated, use it. The best way to limit antibiotic use is to shorten treatment duration. Using 2 different doses just complicates management, makes it more likely that clinicians will make dosing errors, and more likely that patients will “fail” low dose oral antibiotics and be moved unnecessarily to IV. That being said, I don’t manage these infections long term, so perhaps there is a strong evidence based for this practice?

This is a messy trial with a lot of cross-over. The fact that 17% of patients in the oral group switched to IV is a big knock against the idea of oral antibiotics being non-inferior. That being said, many of these switches seem questionable to me. I can understand needing IV on the day of surgery, if you are NPO, but otherwise why does need for surgery change the route of antibiotics? Inability to tolerate oral is a believable pediatric issue, but there is also a lot of subjectivity here. Were they truly unable to tolerate oral? Does that mean they were on IV for the entire duration of treatment, or were they able to tolerate oral for the rest of the week, and just not during the first few days?

On the other hand, they couldn’t get an IV in 9% of the IV group, which probably hints at a lot of harm through repeated attempts at needling these children. 

There were more adverse events in the oral antibiotics group (nausea), but the adverse events they chose to measure were incredibly biased. The list in Table 3 includes stomach pain, nausea, frequent stools, loose stools, and rash. Aside from rash, it looks like they decided to only look for adverse events from oral antibiotics and not IV. What about IV complications? How many times did the IVs fail or come out and have to be restarted? How many times were these kids poked? How do you measure tummy ache and not needle pokes?

Finally, something seems very off with their math to me. Their original power calculation was based on a non-inferiority margin of 10%, and they thought they needed 180 patients. There were no outcomes at all in the first 100 patients, so they adjusted their noninferiority margin down to 5% (which is much harder to demonstrate and needs a much bigger trial), and their new power calculation suggested 184 patients was adequate. That doesn’t make any sense at all. Any statisticians out there care to comment?

I have previously written at length about the many problems with non-inferiority trials. 

Bottom line

This trial has way too many problems to be definitive, but fits with the much larger body of evidence which very consistently says that oral antibiotics are at least as good as IV antibiotics for every infection we have ever thought to test. 

Other FOAMed

Oral antibiotics are equivalent to IV (again) – the OVIVA trial (Li 2019)

Magical thinking in modern medicine: IV antibiotics for cellulitis

You don’t understand non-inferiority trials (and neither do I)

Evidence based medicine is easy

The EBM bibliography

Evidence based medicine resources

EBM deep dives

References

Nielsen AB, Holm M, Lindhard MS, et al. Oral versus intravenous empirical antibiotics in children and adolescents with uncomplicated bone and joint infections: a nationwide, randomised, controlled, non-inferiority trial in Denmark. Lancet Child Adolesc Health. 2024 Jul 15:S2352-4642(24)00133-0. doi: 10.1016/S2352-4642(24)00133-0. Epub ahead of print. PMID: 39025092