In Reply Drs Davidson and Alansari suggest that a difference between high-dose and standard-dose vitamin D supplementation may be found among children with lower baseline serum 25-hydroxyvitamin D concentration at a 90-day end point.
Although not a prespecified analysis, we have conducted subgroup analyses in the interest of hypothesis generation. We considered a negative binomial model with an interaction term between randomization group and baseline serum 25-hydroxyvitamin D concentration using categorical subgroups of less than 30 ng/mL and 30 ng/mL or higher and a 90-day end point (Table).