One piece of wisdom I was given in medical school was to never be the first nor the last to adopt a new treatment. The history of medicine is full of new discoveries that don’t work out as well as the first report. It also is full of long standing dogmas that later were proven false. This balancing act is part of being a professional and being an advocate for your patient. There is science behind this art. Everett Rogers identified innovators, early adopters, and laggards as new ideas are diffused into practice.1
A 2007 French study2 that investigated oral amoxicillin for early-onset group B streptococcal (GBS) disease is one of the few times in the past 3 decades in which I changed my practice based on a single article. It was a large, conclusive study with 222 patients, so it doesn’t need a meta-analysis like American research often requires. The research showed that most of what I had been taught about oral amoxicillin was false. Amoxicillin is absorbed well even at doses above 50 mg/kg per day. It is absorbed reliably by full term neonates, even mildly sick ones. It does adequately cross the blood-brain barrier. The French researchers measured serum levels and proved all this using both scientific principles and through a clinical trial.
I have used this oral protocol (10 days total after 2-3 days IV therapy) on two occasions to treat GBS sepsis when I had informed consent of the parents and buy-in from the primary care pediatrician to be early adopters. I expected the Red Book would update its recommendations. That didn’t happen.
Meanwhile, I have seen other babies kept for 10 days in the hospital for IV therapy with resultant wasted costs (about $20 million/year in the United States) and income loss for the parents. I’ve treated complications and readmissions caused by peripherally inserted central catheter (PICC) line issues. One baby at home got a syringe of gentamicin given as an IV push instead of a normal saline flush. Mistakes happen at home and in the hospital.
Because late-onset GBS can be acquired environmentally, there always will be recurrences. Unless you are practicing defensive medicine, the issue isn’t the rate of recurrence; it is whether the more invasive intervention of prolonged IV therapy reduces that rate. Then balance any measured reduction (which apparently is zero) against the adverse effects of the invasive intervention, such as PICC line infections. This Bayesian decision making is hard for some risk-averse humans to assimilate. (I’m part Borg.)
Coon et al.3 have confirmed, using big data, that prolonged IV therapy of uncomplicated, late-onset GBS bacteremia does not generate a clinically significant benefit. It certainly is possible to sow doubt by asking for proof in a variety of subpopulations. Even in the era of intrapartum antibiotic prophylaxis, which has halved the incidence of GBS disease, GBS disease occurs in about 2,000 babies per year in the United States. However, most are treated in community hospitals and are not included in the database used in this new report. With fewer than 2-3 cases of GBS bacteremia per year per hospital, a multicenter, randomized controlled trial would be an unprecedented undertaking, is ethically problematic, and is not realistically happening soon. So these observational data, skillfully acquired and analyzed, are and will remain the best available data.
This new article is in the context of multiple articles over the past decade that have disproven the myth of the superiority of IV therapy. Given the known risks and costs of PICC lines and prolonged IV therapy, the default should be, absent a credible rationale to the contrary, that oral therapy at home is better.
Coon et al. show that, by 2015, 5 of 49 children’s hospitals (10%) were early adopters and had already made the switch to mostly using short treatment courses for uncomplicated GBS bacteremia; 14 of 49 (29%) hadn’t changed at all from the obsolete Red Book recommendation. Given this new analysis, what are you laggards4 waiting for?
Dr. Powell is a pediatric hospitalist and clinical ethics consultant living in St. Louis. Email him at firstname.lastname@example.org.
1.5th ed. (New York: Free Press, 2003).