This paper employs an Extreme Value Theory framework to investigate the existence of contagion between European and US banks. The fact that many regulators have no detailed data sets about interbank cross-exposures raises the necessity of finding market-based indicators in order to analyze the effects of crises and to quantify the risk of contagion. The Distance-to-default (DD) measure is being employed as an indicator of banks' soundness. Focusing on the negative tail of the daily percentage changes of the DD, a country-specific indicator variable labeled "Coexceedances" is built measuring the number of banks simultaneously experiencing a large shock on a given day. Based on a multinomial logit model, for each country the probability of observing several banks in the tail is estimated. Controlling for common factors and including foreign countries' lagged coexceedances allows to interpret significant coefficients of foreign lagged coexceedances as contagion. The main finding is that there is significant bi-lateral contagion between European and US banks. Furthermore the existence of contagion between European banks is verified by the underlying data set.