Hemangioma Treatment: Should We Just Skip Controlled Trials?

Source: Price CJ, Lattouf C, Baum B, et al. Propranolol vs corticosteroids for infantile hemangiomas. Arch Dermatol. 2011;147(12):1371-1376; doi:10.1001/archdermatol.2011.203. See AAP Grand Rounds commentary by Dr. Teresa Wright (subscription required).

PICO
Question: Among patients with problematic infantile hemangiomas, does treatment with propranolol compared to oral corticosteroids improve clinical outcomes and decrease adverse effects?
Question type: Treatment
Study design: Multicenter retrospective chart review

I did a little background checking on the term, "jumping on the bandwagon," because I think the authors of this study come dangerously close to premature jumping.

As often is the case, Wikipedia is a decent start for background information. Their posting on bandwagon effect describes the term as originating in 1848, when a circus clown turned politician (this blog is apolitical, but boy am I tempted here!) used his circus bandwagon in his political campaign. Now, the term applies to any phenomenon that tends to increase its adoption in proportion to the number of people who have already adopted it.

I have some trouble accepting the authors' final sentence in the article: "...we believe that propranolol therapy is superior to traditional first line oral corticosteriods in the treatment of [infantile hemangioma], and we propose that propranolol be considered a first-line agent given its safety and efficacy..." My real problem is that their study is a retrospective chart review, with all the accompanying bias that is impossible to estimate in that type of study design. Outcomes included non-standardized monitoring of response to therapy and relapses, use of imaging data, and even use of alternative treatments such as topical or intralesional steroids and laser therapy (22 patients of the original 139 were excluded because of these alternative treatments). It would have been helpful to have more details for these patients, particularly as to whether they might be considered propranolol failures.

On the other hand, it's probably impossible to perform a double-blinded randomized prospective trial of propranolol versus steroid therapy, since the presence of cushingoid features in steroid recipients clearly would "unblind" both investigators and parents. Still, like the editorial comment at the bottom of the AAP GR summary, I'd favor a prospective randomized trial before jumping on the propranolol bandwagon.