Background: Atypical interoception has been observed across multiple mental health conditions, including anxiety disorders and depression. Evidence suggests that not only pathological anxiety, but also heightened levels of state anxiety and stress are associated with interoceptive functioning. This study aimed to investigate the effects of the recent Coronavirus SARS-CoV-2 pandemic on self-reported interoception and mental health, and their relationship. Methods: Self-report measures of interoceptive attention and accuracy, anxiety, stress and depression taken during the pandemic (at three time points) were compared to the same measures taken from comparable samples prior to the pandemic. In the sample collected during the pandemic, the relationship between interoceptive and mental health measures and focus on COVID-19-related news and information, propensity to take objective measures of COVID-19 symptoms, and subjective beliefs concerning COVID-19 symptoms was assessed. Finally, a cross-lagged panel model (CLPM) was used to test directional relationships between self-reported interoceptive and mental health measures across three time points. Results: Higher self-reported anxiety was associated with a) increased self-reported attention to bodily signals, b) increased focus on COVID-19-related news and information, c) propensity to take objective measures of COVID-19 symptoms, and d) reduced self-reported interoceptive accuracy for bodily signals participants believed were associated with COVID-19. The CLPM revealed a mutual and comparable directional effect from T1 to T2 between interoceptive attention and measures of mental health. Conclusions: Implications of these findings are discussed in the light of existing models and newly proposed accounts of the relationship between interoception and mental health.