Still ignoring post-operative deaths

Last year ripe-tomato.org commented on a neonatal circumcision trial conducted by US researchers in Botswana that had been registered (here) as a three arm trial, but reported with only two (click here). As the authors explained (here):

“The trial was originally designed to evaluate only the Plastibell and the Mogen clamp among 300 newborn boys (randomized 1:1 to circumcision method, 150 per arm). We learned of the availability of the newer AccuCirc device only after initiation of the original trial. […] we added the AccuCirc arm as an amendment to our original study protocol, with a plan to circumcise an additional 150 neonates by AccuCirc (leading to a total of 450 neonates circumcised in the 3 arms, as reflected on ClinicalTrials.gov). Because the AccuCirc arm started enrolling only after the original 2 randomized Plastibell and Mogen arms had completed enrollment, the AccuCirc study results were not included in the manuscript submitted to JAIDS. […] a manuscript summarizing the findings from the AccuCirc arm is in preparation.”

The missing AccuCirc arm has just appeared. Plank Accucirc 2014 paper. From a “convenience sample” of 151 infants, all but one were seen at 7 days. The parents of the missing infant, contacted by phone, reported “no adverse events or complications”.  Five infants had mild bleeding, one severe bleeding associated with vitamin K deficiency and there was one unconfirmed infection. No deaths were reported.

Some readers might question if such high follow-up is plausible. My quibble is with the authors citing their previous trial as showing “a very low rate of serious adverse events associate with EIMC” without mentioning the three deaths!

But I guess that’s what you’d expect from people who classify an un-investigated death in a previously healthy boy less that 24 hours after circumcision as “extremely unlikely” to be procedure-related! Read more here.

Jim Thornton