I suspect the random variation between the two would account for far more difference than the casualties directly attributable to method of destruction, though.

Unless you mean destroying the transformers with blockbuster-size HE, or dirty nuclear devices?

I guess my point is that the number of people who would die from sustained, widespread infrastructure disruption would be huge, and not greatly changed by casualties around the blast site if you decided to blow stuff up literally instead of just metaphorically.