That's not what the best data available would suggest. Every study that's looked at the correlation between USMLE and COMLEX has shown that exam performance tends to correlate (with some scatter, of course), and that the minimum pass rate for COMLEX is somewhat below that of USMLE.
For example:
The Use of COMLEX-USA and USMLE for Residency Applicant Selection This study looked at three different schools. They only performed a pearson's correlation and not a regression, but you can look at their image and see the results. If you draw a line down the middle of their elipse (on the graph), you'll see that the minimum pass of USMLE 194 is well above the minimum pass of COMLEX (400). They also include a 2x2 table of pass/fail for the two exams, show moderate correlation (a stretch, at best) with the majority of the discordance a pass on COMLEX and a fail on USMLE.
Reliability and validity of conversion formulas between comprehensive osteopathic medical licensing examination of the United States level 1 and Un... - PubMed - NCBI This article looks at only a single school but over a large number of years. You can take their regression formula and plug in a USMLE of 194, and you'll get a COMLEX of 440 or so.
There are many more, all show the same thing. They all are hampered by similar problems -- not all DO's take both exams, and difficulty getting a non-biased sample (either all one school, or all applying to one residency program, etc). Proponents will point out that the pass rate on S1 is basically equal for MD and DO now, clearly improved over the years. But if those that score poorly on the COMLEX are more likely to forgo the USMLE, then that statistic may be biased upwards.
So I don't expect COMLEX to go anywhere, as I think schools would be very concerned about ultimate pass rates on USMLE alone.