Microvascular decompression (MVD) is associated with low mortality and morbidity rates at specialized centers, but many MVD procedures are performed outside such centers. We studied short-term end points after MVD in a national hospital discharge database sample. A retrospective cohort study was performed by using the Nationwide Inpatient Sample, 1996 to 2000. The sample included 1326 MVD procedures for treatment of trigeminal neuralgia, 237 for treatment of hemifacial spasm, and 27 for treatment of glossopharyngeal neuralgia, performed at 305 hospitals by 277 identified surgeons. The mortality rate was 0.3%, and the rate of discharge other than to home was 3.8%. Neurological complications were coded in 1.7% of cases, hematomas in 0.5%, and facial palsies in 0.6%, with 0.4% of patients requiring ventriculostomies and 0.7% postoperative ventilation. Trigeminal nerve section was also coded for 3.4% of patients with trigeminal neuralgia, more commonly among older patients (P = 0.08), among female patients (P = 0.03), and at teaching hospitals (P = 0.02). The median annual caseloads were 5 cases per hospital (range, 1-195 cases) and 3 cases per surgeon (range, 1-107 cases). With adjustment for age, sex, race, primary insurance, diagnosis (trigeminal neuralgia versus hemifacial spasm versus glossopharyngeal neuralgia), geographic region, admission type and source, and medical comorbidities, outcomes at discharge were superior at higher-volume hospitals (P = 0.006) and with higher-volume surgeons (P = 0.02). Complications were less frequent after surgery performed at high-volume hospitals (P = 0.04) or by high-volume surgeons (P = 0.01). The rate of discharge other than to home was 5.1% for the lowest-volume-quartile hospitals, compared with 1.6% for the highest-volume-quartile hospitals. Volume and mortality rate were not significantly related, but three of the four deaths in the series followed procedures performed by surgeons who had performed only one MVD procedure that year. Length of stay (median, 3 d) and hospital volume were not significantly related. Hospital charges were slightly higher at higher-volume hospitals (P = 0.007). Although most MVD procedures in the United States are performed at low-volume centers, mortality rates remain low. Morbidity rates are significantly lower at high-volume hospitals and with high-volume surgeons.