This is why most GT program starts at third or fourth grade. I seem to recall for all the kids tested gifted at 4 or 5 for K entry, only 30 or 40% will test gifted at 8 or 9, and a lot of truly gifted kids won't test well at a younger age due to maturity issues.
Of course, the higher the number at 4, the less likely the child is going to slow down later. But at the margin, the false positive/negative rates are high.
This is the dilemma. Early testing is unreliable. But at these young ages, the differences in ability, and the differences in educational needs, are definitely there. These early grades need to let kids move at their own pace, and not try to keep them all in lockstep.