The PN regimens differed between studies and even within studies. There was no conflict of interest in any of the 4 studies included in the meta-analysis; they were all supported by national granting agencies or institutional funds. In all 4 studies, the randomization process was performed in the participating centers using consecutively numbered sealed envelopes provided at the beginning of the study