We compare published solutions for the rotation angle W(t) which describes the location of the prime meridian of Mars with respect to the ICRF equator in IAU recommendations. If the model for W includes a very long period term, we transform it into a quadratic polynomial with updated epoch value and rate, resulting in a difference in the mean epoch value up to 200 km. The mean and true epoch rotation angles are about 800 mas (13 m) apart in J2000 and should not be confused with each other in order to accurately locate the prime meridian. We identify two groups of radio-science solutions for W, which can be distinguished by the prime meridian location they used as a priori that differ from each other by about 100 m at J2000.