This study examined attrition in a multisite clinical trial for panic disorder. Of 379 eligible patients, 19 refused treatment (5% Refusal Rate), 104 discontinued treatment prematurely (19% Dropout Rate) or were withdrawn by the investigators (8% Withdrawal Rate), and 256 completed the treatment (68% Completion Rate). Logistic regression was used to examine 5 domains theorized to be related to attrition (e.g., diagnostic severity, treatment factors). Few variables were associated with increased odds of attrition at padj < 0.004. Younger age was the only independent predictor of attrition in the demographic factor model. Diagnostic severity and comorbidity, panic disorder symptom severity, treatment factors, and therapist factors were unrelated to study attrition. Patient dropout was highest after treatment sessions that targeted interoceptive and situational exposure exercises. Findings suggest that attrition may not strongly threaten the validity of results from treatment outcome studies.