To investigate whether Hurricane Katrina's landfall in August 2005 resulted in excess mortality, we conducted a cohort study of patients who started dialysis between January 2003 and late August 2005 and who received treatment at 94 Katrina-affected clinics in the area. Survival, regardless of patient location after the storm, was followed through February 2006. In adjusted Cox proportional hazards models, Hurricane Katrina (time-varying indicator) was not significantly associated with mortality risk for patients from regions of the Gulf Coast affected by Katrina or those from a subset of 40 New Orleans clinics. Subgroup analyses indicated no significant increased mortality risk by race, income status, or dialysis modality. Sensitivity analyses indicated no significant increased mortality risk for patients from clinics closed for 10 days or longer, patients in their first 90 days of dialysis, or patients not evacuated from the affected areas. Patients remaining in the New Orleans area may have been more vulnerable due to age and comorbidities; however, the change in their mortality risk in the month following the storm was not statistically significant. We suggest that disaster-related education for patients must be ongoing, and that each disaster may present a different set of circumstances and challenges that will require unanticipated response efforts.