Dental anxiety (DA) arises from exogenous learning experiences and endogenous vulnerabilities, yet little is known about how these factors combine within highly anxious adults. The aim was to identify empirical subtypes of high DA based on endo/exogenous factors and compare them according to self-reported oral health, care-seeking behavior, and preferred dental treatment conditions.

A cross-sectional online survey of 399 Japanese adults (mean ± S.D. age = 43.5 ± 13.1 years; 50% women) who rated “very frightening” on the Japanese Single Dental Anxiety Question. Seven predictors were entered into a two-step cluster analysis: four exogenous variables (distressing dental and medical experiences, family DA, exposure to frightening dentists in the media) and three endogenous variables (fear of pain, pain catastrophizing, sensory-processing sensitivity). Demographics, Seattle DA classification, dental attendance pattern, self-reported oral health, and preferred treatment conditions (cognitive-behavioral therapy, conscious sedation, or general anesthesia) were compared between clusters using chi-square and Mann–Whitney U tests.

The optimal solution consisted of two clusters (silhouette = 0.30, maximum/ minimum ratio = 1.31). All participants in Cluster 1 (n = 173) had experienced distressing dental treatment and showed higher pain catastrophizing (p < 0.001). No one in Cluster 2 (n = 226) reported traumatic dental events and had lower catastrophizing, fewer regular check-ups (12% vs. 24%), a higher proportion who had never visited a dentist (14% vs. 0%), and less interest in DA treatment compared to Cluster 1.

Two partially overlapping DA phenotypes emerged: C1 as “Trauma & Catastrophizing Compounded C1/ Still-Considers” and C2 as “Latent-Trait Driven C2/ Extreme Avoiders.” Tailored strategies that combine trauma-informed exposure for the former and sensory-adapted engagement for the latter may enhance the uptake of oral healthcare. Broader endogenous indicators should be added in future work to refine subtype boundaries.