In Digital Fusion, I've dragged the rendered image into a new composition, created a copy of the Loader node, and on the copy's Format tab, set the Red, Green and Blue channels to be gotten from the Reflection (Refl.*) channels, which exrTrader Layered into the image file. (The "Final Render" is on the left, the Reflection Buffer is on the right.)

I want to bump-up the intensity of the reflections in my image, so I increase the Gain on the Reflection channels with a Color Corrector node, and then Merge it over my original, Rendered Image using the Screen mode. (I find that Screen, Lighten or Color Dodge will usually give me the effect I'm looking for – they each give me slightly different "flavors" of increasing the light areas of an image with the light areas of another.)

Next, because I didn't tell exrTrader to invert LightWave's Depth (Z-Buffer) Channel, I need to bring in a Channel Boolean node, telling it to "Do Nothing" to all Channels except the Z Buffer, and to it, mathematically invert it. (You see on the left, the original Z-Buffer as it comes out of LightWave, and on the right, you see it as Fusion needs to see it, with the PIXEls farther away from the camera being darker than those closer to it.)