This came up where I work as well. The policy that came out in the end was to continue to use IV valium for DT's. IM valium has not been shown to be effective when treating DT's, and lorazepam is not the preferred drug for DT's due to it's half-life, as well as studies of the outcomes of DT patients medicated with lorazepam vs. diazepam.
Our policy is to dilute it. While it may form a precipitate, that's actually OK: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2516026/ http://www.ncbi.nlm.nih.gov/pmc/arti...00131-0020.pdf http://www.ncbi.nlm.nih.gov/pubmed/7294034
IV valium has a low solubility in NS and other IV solutions, which is why the IV solution is 40% propylene glycol. Once it meets blood however, it re-disolves. The reason why we our policy is to dilute it is that undiluted valium, with it's 40% antifreeze (propylene glycol), has a significantly higher risk of causing cardiac adverse effects (asystole).