We use detailed administrative records to show that high household leverage increases mental health fragility, with persistent negative economic effects. After adverse life events, e.g. heart attacks or job losses, individuals with higher ex ante leverage experience larger increases in mental health problems. The effects are long-lasting and stronger in times of financial crisis. Parallel pre-trends, robustness to non-parametric controls, and IV estimation suggest the results are not driven by confounding unobservables. High leverage is also associated with worse long-run earnings dynamics at the time when loan arrears and mental health problems emerge, suggesting tenacious scarring effects of leverage.