You’re correct that the advantages of top aligned labels that Penzo found (in your reference) only apply when the user is moving from the top to bottom, looking at each field in order. That may be the case for a printed report –you need to find out from studying your users. Do they read the whole thing like a newspaper story?
However, for most reports, I suspect users look for specific fields or values, and scan and skip around the page. For that, you probably want to have left justified field values and labels in two parallel columns so users can more easily scan for a particular field or value. The problem of associating the labels with the values over wider separations can be ameliorated by using leader dots or zebra stripes. It may also be best to use abbreviations or multi-line labels to try to keep the values consistently close to the labels.
In either case, you probably can make best use of paper by laying the fields out in two or more columns like a physical magazine. Lay out the fields so the user reads down a column before shifting to the next column. Use appropriate graphic design to cue the user to read downward first.
Of course, the use of multiple columns assumes your users do indeed want/need printouts of these reports. In many organizations it makes more sense to provide electronic reports (e.g., in HTML or something more interactive) for on-screen viewing. Sometimes I think organizations are still printing out and distributing paper reports just because “that’s the way we’ve always done it.” Or they distribute the reports as .pdfs files (e.g., by intranet), without thinking maybe a different format should be used for better on-screen viewing.