We have performed a careful investigation of the 74 candidate variable stars presented by Safonova & Stalin. For this purpose we used our data base of imaging and light curves from Arellano Ferro et al. We find that two candidates are known variable stars, eight candidates were discovered first by Arellano Ferro et al., but would not have been known to Safonova & Stalin at the time of their paper submission, while four candidates are new variables. Three of the new variables are SX Phe type and one is a semiregular (SR) type red giant variable. We also tentatively confirm the presence of true variability in two other candidates, and we are unable to investigate another four candidates because they are not in our data base. However, we find that the remaining 54 candidate variable stars are spurious detections where systematic trends in the light curves have been mistaken for true variability. We believe that the erroneous detections are caused by the adoption of a very low detection threshold used to identify these candidates.