My baby has eczema.  Eczema is an inflammatory condition of the skin characterized by redness, itching and oozing vesicular lesions which become scaly, crusted, or hardened (and thank you for that, Merriam-Webster!).  What this means for my baby is that his face, trunk and upper arms are spotted with itchy red sores, which obviously distress him, lead to poor sleep (for him and therefore everybody), and cause him to scratch so vigorously as to routinely bloody himself.  The eczema has also put a stop to what would surely otherwise have been a promising career as a plus-sized baby model.

Eczema in infants is usually managed by keeping the affected areas clean and moisturized, and with the use of topical corticosteroids as needed and sometimes antihistamines to reduce itching.  Babies with eczema are also prone to bacterial skin infections and these are usually managed with oral and/or topical antibiotics when they occur.  There's also an impressive array of specialized products available, although for the most part these seem to be indistinguishable from non-specialized counterparts, marketing aside.  Still, my wife and I have an acute interest in finding effective ways to lessen our baby's discomfort, and so she has spent a considerable amount of time doing research on the web, trying to find credible recommendations for specific moisturizers, shampoos, etc. from parents in situations similar to our own.

In the course of this research, she stumbled on this conversation at babycenter.com, in which a distressed parent is asking for advice on using a diluted bleach bath as a treatment for eczema.  The first few (and many subsequent) responders had the same reaction my wife and I had and which, perhaps, you are now sharing: that's crazy, stupid, and/or crazy stupid.  I mean, bleach + baby with sensitive skin = not a real equation, but probably a mistake nonetheless.  But wait, says commenter deb70b, it's science!  And she thoughtfully provides links to several credible news sources discussing diluted bleach baths as an effective treatment for eczema, including the New York Times, Time Magazine and Babycenter itself (reprinting a HealthDay News wire story).  Google and you'll find more of the same.  Each of these articles is based on a study published in the May 5th, 2009 issue of the journal Pediatrics, which shows that diluted bleach baths are an effective treatment for eczema.

Except it doesn't.  When I went to the Pediatrics archives to look up the study, I didn't even see it in a first pass of the table of contents, since I was looking for something entitled "Bleach Baths as a Treatment for Infantile Eczema," or some such something.  The article in question is instead called, "Treatment of Staphylococcus aureus Colonization in Atopic Dermatitis Decreases Disease Severity."  (Atopic dermatitis is the chronic condition which underlies most cases of childhood eczema.)  The journal is behind a pay wall, but the abstract is publicly accessible, and I'm going to go ahead and add some emphasis to that sucker:

METHODS. A randomized, investigator-blinded, placebo-controlled study was conducted with 31 patients, 6 months to 17 years of age, with moderate to severe atopic dermatitis and clinical signs of secondary bacterial infections. All patients received orally administered cephalexin for 14 days and were assigned randomly to receive intranasal mupirocin ointment treatment and sodium hypochlorite (bleach) baths (treatment arm) or intranasal petrolatum ointment treatment and plain water baths (placebo arm) for 3 months. The primary outcome measure was the Eczema Area and Severity Index score.

Now, I think this was a reasonably well-designed preliminary (!) trial, with one exception to be noted below, and as far as I can tell it was conducted properly.  But it doesn't remotely imply what any of the media reports claim it implies.  Even though those reports extensively quote the last author of the study, she herself makes claims that are clearly unsupported by the data.  (Traditionally, the last author of a manuscript in a medical journal is the person in whose lab the research took place or who was responsible for the research funding; the first author merely does all the work.)

Let's take the points at which the actual study and the study as imagined by the New York Times et al. diverge one at a time:

1.       31 patients.  Not very many, is it? In the terminology of clinical trialists, a study this size would usually be considered a Phase I study, and would be intended to show feasibility and preliminary evidence of the safety of a proposed treatment, and perhaps to look for a maximum tolerated dose.  Anyone trying to pass off a study this size as providing conclusive evidence of efficacy would be laughed out of science court.  There's nothing wrong with preliminary studies.  The world needs preliminary studies - lots of them - to weed out bad treatments and identify potentially effective ones.  But it is grossly irresponsible to publicize the results of such a study without emphasizing its preliminary nature, especially considering the "try this at home!" nature of this particular treatment (more on this below).  The Times didn't even provide the sample size, much less contextualize it.

2.       A related specific objection to the Babycenter report and its like which are recommending bleach baths as treatments for babies: This was 31 patients, aged 6 months to 17 years, divided into two groups.  It stands to reason that not too many infants were actually given the bleach bath treatment, right?  In fact, none were.  Table 1 of the paper (behind the pay wall) reports demographic data for the study sample; the minimum age in the treatment (i.e. bleach + mupirocin) group was 2.1 years.  (The minimum age in the placebo group was 0.7 years, so at least one infant was, um, not given bleach baths as part of the study.)  It's pretty clear why this is a problem:  One of the rules of clinical trials (and of any research, really) is that a study's results generalize to the population from which the study sample was taken.  You can't really say anything about bleach baths for babies without actually putting some babies in bleach baths.  And this segues nicely into the next reason those media reports were misleading.

3.       This study doesn't say anything about babies with just plain old eczema.  It says something (preliminary!) about babies with eczema and "clinical signs of secondary bacterial infections."  Namely, it says that maybe you can use bleach baths as part of the treatment regimen for the bacterial infections, and that treating the bacterial infections might help reduce the eczema symptoms.  Our baby, for instance, hasn't had any secondary bacterial infections; just the eczema.

4.       The final major failing, and the flaw in the study design I mentioned earlier, is this: If the authors had wanted to study the effects of bleach baths, they should have had two study groups, one that received bleach baths, and one that did not receive bleach baths but was otherwise treated exactly the same.  Instead of this, the treatment group received both bleach baths and intranasal mupirocin (a topical antibiotic).  The placebo group received intranasal petrolatum (Vaseline) and plain water baths instead.  So any perceived benefit for the treatment group could be due to the bleach, the mupirocin, or both.  Or even to detrimental effects of the plain water baths used in the placebo group (some dermatologists recommend limiting bathing for infants with eczema).

Diluted bleach baths as described in this study are unlikely to be actually unsafe, even for babies with eczema.  After all, the concentrations of chlorine here compare favorably to those found in swimming pools.  But it's nevertheless extremely irresponsible to advise parents to try something like this without much stronger evidence that the benefits outweigh the risks.  If thousands of parents try this, it's pretty much a guarantee that a certain number of them will screw up the bleach ratio, potentially harming their children.  For instance, it's not hard to imagine someone adding the recommended half-cup of bleach not to a standard 40 gallon tub, but instead to a typical 3 - 5 gallon infant tub.

While it's fun (and, I hope, useful) to trash the mainstream media for screwing up a specific health or science news story, the truth is that this is more the norm than the exception.  In future articles, I hope to discuss some proposals to improve this situation, including prescriptions for semi-standardized reporting of clinical trials (always reporting the sample size, for instance).  But nothing can replace constant vigilance from the critical thinking community.  If you see a fishy-seeming medical news story, leave a comment about it, or come discuss it on the ILS forums!

Discuss this article on our forums