The coronavirus disease 2019 (COVID-19) pandemic led to various government-imposed limitations on social interaction and strict home confinement. Such involuntary social-distancing policies can exacerbate feelings of loneliness and alter emotional well-being. Dysregulation of the hypothalamic-pituitary-adrenocortical (HPA) axis is a potential mechanism for loneliness' deleterious health effects. In this study, we explored whether pre-pandemic diurnal cortisol output (AUC(G)), a measure of HPA axis function, may predict the propensity to changes in loneliness during long-term COVID-19 home confinement and if extraversion would moderate this relationship. This association has been explored by analysing the impact of COVID-19 pandemic and strict home confinement on social and emotional loneliness in 45 Spanish young adults. Diurnal cortisol levels were measured from five saliva samples obtained across a day just before the pandemic, and data about participants' perceived loneliness, empathic state, extraversion, and prospective volunteering were obtained both before and during the confinement. Participants' social and family loneliness increased during long-term strict home confinement, while prospective volunteering tendencies and extraversion decreased. Importantly, after adjusting for relevant confounders, moderation analyses revealed that in young adults with high pre-pandemic extraversion, a higher AUC(G) predicted a larger increase in social loneliness during confinement, while in individuals with low extraversion, AUC(G) was negatively related to change in loneliness. Our findings highlight the utility of pre-pandemic diurnal cortisol output in predicting the social impact of COVID-19 home confinement, presenting this hormone as a potential biomarker for a priori identification of at-risk groups during public health crises.