BACKGROUND: Whether oxytocin functions as a stress hormone in older age is unknown. We investigated levels and the perceived stress of an adverse life event in an older population-representative sample and considered the effect of a secure/insecure attachment style on this association. METHODS: Non-fasting plasma oxytocin was measured from 952 participants (65-90 years) of the cross sectional KORA-Age study. The psychological impact of an adverse life event was assessed based on the Psychosocial Stress Questionnaire. Attachment style was determined by the Relationship-Specific Attachment Scales for Adults. Linear regression models of oxytocin, stratified for attachment style, were controlled for age, sex, and further for alcohol, smoking, and physical activity. Adjusted least squares means of oxytocin were calculated. RESULTS: Oxytocin levels did not differ between men and women (mean, 95% confidence interval (CI), 321 (277-365) and 309 (272-345)pg/ml, respectively). Oxytocin levels were positively associated with the experience of an adverse event (n=273, 29%) versus no event (n=679, 71%), in securely attached (β estimate=0.17, standard error (SE)=0.08, P value=0.03) but not in insecure participants (-0.10, 0.09), P=0.28). Oxytocin was positively associated with diminished stress among securely attached participant (event with little suffering: β=0.35. SE=0.12, great suffering: β=0.15. SE=0.14, severe suffering: β=0.03. SE=0.12). Among participants who reported minimal suffering, insecure individuals had lower oxytocin (adjusted mean, 95%CI: 172, 127-216pg/ml) than securely attached individuals (279, 222-352pg/ml, P=0.006). CONCLUSIONS: These epidemiologic data support the hypothesis that oxytocin may have an attenuating effect on perceived stress due to adverse life events in old age. The conditional role of attachment style in stress-induced endogenous oxytocin production is highlighted.