The best way to make this conversion is to do the experiment yourself under your own assay conditions. You need a way of measuring the micromoles of product formed. Then you can compare the rate you measure to the number of NIH units that went into the reaction.