This pooled analysis evaluated potential predictive abilities of baseline demographic factors, psychiatric history, and DSM-IV diagnostic criteria for short- and long-term outcome after treatment with venlafaxine extended release (XR) or placebo in patients with generalized anxiety disorder (GAD). Pooled data from 1,839 patients in five placebo-controlled studies of venlafaxine XR for GAD were analyzed by logistic regression. Odds ratios (ORs) were used to quantify pretreatment factors' abilities to predict response (50% reduction, baseline Hamilton Rating Scale for Anxiety [HAM-A] severity) and remission (total HAM-A score </=7) following venlafaxine XR or placebo treatment. All analyzed factors showed statistically significant outcome associations after 8 or 24 weeks of treatment, or both, in placebo- or venlafaxine-XR-treated patients or both. Substance abuse history, DSM-IV diagnostic criteria of sleep disturbances, difficulty concentrating, and restlessness had the strongest associations with outcome. Sleep disturbances predicted significant positive response in both groups, but more so in the placebo group. Restlessness consistently predicted poor response and lack of remission with either treatment; difficulty concentrating predicted short-term remission with placebo only. Substance abuse history predicted positive outcomes with placebo only. Sex, age, depression history, panic disorder history, prior benzodiazepine and nonbenzodiazepine use, being easily fatigued, muscle tension, and irritability were modestly outcome-predictive or showed treatment condition interactions. In the largest pooled analysis to date, pretreatment factors were associated with treatment outcome in patients with GAD receiving venlafaxine XR or placebo. The strongest trends emerged for history of substance abuse or dependence and symptoms of restlessness, sleep disturbance, and difficulty concentrating.