Processing threatening stimuli is thought to have priority over other types of content. Current studies have shown that neutral stimuli that reliably predict the appearance of threatening content, can elicit anticipatory attentional bias. Using this cued visual probe task (CVPT) paradigm can enhance the reliability of measuring this bias because in this setup the bias is triggered by the category of the content (threat), rather than the individual characteristics of the threatening stimuli used. In the present study, we investigated whether this anticipatory effect occurs to specific, medical fear-related (medical equipment, injuries, blood) stimuli as threatening content using neutral stimuli as predictors. In two experiments, we used two CVPT paradigms with different designs. In Experiment 1 (N=33), pairs of predictive cues, threatening and neutral images, and the probe appeared in a diagonal arrangement on the screen (similarly to the classical dot-probe task). In Experiment 2 (N=51), cues and the content they predicted were presented in the middle of the screen (individually), while the probe was still presented in an off-central position. In both setups, the task was to indicate where the target probe was presented by pressing the corresponding button. In Experiment 3, we used eye-tracking (N=29) to replicate and extend Experiment1 and 2. Participants tended to avoid the location of the predictive stimulus associated with threatening content, resulting in slower (Exp1) and faster (Exp2) reaction times in solving the task. The results of the eye-tracking experiment help to place our results in the context of attentional biases.