Treatment resistant OCD subjects, defined as those patients who undergo an adequate trial of SRI (clomipramine or SSRI) and do not respond or show unsatisfactory results, account for 40-50% of all patients. Once the appropriateness of the trial has been assessed, several options exist for the clinicians. If clomipramine or citalopram have been used, an appropriate strategy consists in giving the same drug intravenously. Double-blind studies exist on the efficacy of clomipramine IV, while data are missing for citalopram. Another option that should be considered first, although data are scarce, is the addition of a cognitive behavioral therapy, when available, in the forms of exposure and response prevention. When such options are not suitable or available, augmentation of the ongoing SRI with another compound represents the preferable strategy. Double-blind, placebo-controlled studies have shown the efficacy of adding pindolol (7.5 mg/d), risperidone (2 mg/d) and olanzapine (5-10 mg/d). Other agents have been proposed, but data emerging from double-blind studies were negative or contradictory. Another option available is switching from CMI to SSRI, or vice versa, or from SSRI to SSRI. Data regarding such treatment strategy, however, are highly preliminary, based on a couple of open label reports and on studies performed in treatment resistant depression. An unresolved question is whether augmentation should be preferred to switching. No data exist in OCD; a practical approach would suggest augmentation first, considering that response should be obtained faster than by switching compound. When all the available and effective strategies prove uneffective, clinicians should consider switching the patient to other compounds in monotherapy, such as venlafaxine, sumatriptan, inositol, although research is strongly needed before conclusions on the efficacy of such compounds can be drawn.