sults suggest that combined treatment, at least as applied simultaneously, has not proven to be a panacea (Foa, Franklin, & Moser, 2002; Franklin & Foa, 2002). In trials directly comparing serotonin reuptake inhibitors (SRIs) and CBT (including EX/RP), outcome has been influenced by the design of the study, the SRI dosage, and the quality and the quantity of the EX/RP delivered. When intensive EX/RP (e.g., 15 sessions delivered five times per week for 3 weeks) was compared to adequately dosed clomipramine over 12 weeks, arguably the two most efficacious OCD treatments, the two treatments were both superior to placebo and EX/RP was superior in most analyses (Foa et al., 2005). Further, when results from this study using an intensive CBT approach are compared with a study employing weekly CBT with and without concomitant medication (Cottraux et al., 1990), intensive CBT alone appears to fare better than weekly CBT alone, with no differences evident in the combined treatment groups (Foa, Franklin, & Moser, 2002). It is important to note that response prevention in the study by Foa et al. (2005) appears to have been more strictly applied than that in the Cottraux et al. study (1990), and thus the effects of the visit schedule cannot be isolated in this comparison. Nevertheless, perhaps the addition of medication may enhance outcome when EX/RP is not implemented intensively or when variations are made to the EX/RP protocol.

Children and adolescents with OCD typically present with both obsessions and compulsions, although the youngest sufferers may have difficulty articulating their obsessions. As is the case with adults, the cardinal feature of OCD in youth is neutralizing: when the patient describes anxiety-inducing thoughts or images and attempts to relieve this anxiety or reduce the chances that feared consequences would occur by performing some overt or covert neutralizing behavior, the OCD diagnosis should be considered. Functional impairment is required for diagnosis as well, as subclinical obsessions and compulsions are probably ubiquitous. Insight into the senselessness of obsessional concerns is not required for diagnosis in children and adolescents, and, as with adults, probably exists along a continuum from complete awareness of their senselessness to no insight (Foa, Hearst-Ikeda, & Perry, 1995). In general pediatric OCD is formally similar to OCD in adults, yet the content of obsessions and compulsions is likely to be influenced by developmental factors. For example, younger children are generally more “magical” in their thinking and thus may have more superstitious OCD symptoms (e.g., “If I don't retrace my steps then something really bad will happen to my little sister”). As with adults, some pediatric OCD patients are able to identify feared consequences of not ritualizing (e.g., books will be stolen if the locker is not checked), whereas others experience anxiety and distress in the absence of articulated consequences. Further, although the logic of some patients' feared consequences is shared by many in their culture (e.g., contracting disease via direct contact with a public toilet seat), other patients' fears are extremely unusual (e.g., losing their essence by discarding trash that has touched them). It is important to recognize that bizarre content does not necessarily preclude a diagnosis of OCD and that patients with such unusual fears may also be responsive to CBT (Franklin, Tolin, March, & Foa, 2001).

A subgroup of children with pediatric onset of either OCD or a tic disorder has been described by the term PANDAS (Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal infection). These children have an abrupt onset of symptoms after a group A β-hemolytic streptococcal infection (GABHS), and their course of illness is characterized by dramatic acute worsening of symptoms with periods of remission (Swedo 1994, Swedo et al., 1998). The PANDAS subgroup is defined by five clinical characteristics: (1) presence of OCD and/or a tic disorder; (2) prepubertal symptom onset; (3) dramatic onset and acute exacerbations with an episodic course of symptom severity; (4) temporal association between symptom exacerbations and GABHS infections; and (5) associated neurological abnormalities (e.g., choreiform movements) (Swedo et al., 1998). Identification of this subtype is important because the symp

toms may require a different assessment and treatment. In a child who has an acute onset of OCD and/or tics or has had a dramatic deterioration, medical illnesses (including seemingly benign upper respiratory infections) in the prior months should be carefully considered. A throat culture, antistreptolysin O (ASO) titer, and anti-DNaseB streptococcal titer may help to diagnose such an infection, even in the absence of clinical symptoms of pharyngitis (Murphy & Pichichero, 2002; Swedo et al., 1998).

Expert Consensus Guidelines (March, Francis, Kahn, & Carpenter, 1997) and the American Academy of Child and Adolescent Psychiatry (AACAP) Practice Parameters for OCD (King et al., 1998) consider CBT, specifically the features of EX/RP, an important intervention, and recommend starting with CBT or CBT plus an SSRI, depending on severity and comorbidity. Exposure and response prevention involves therapist-assisted in vivo exposure to feared situations, imaginal exposure to feared “disasters,” and instructions to refrain from rituals and avoidance behaviors. In the treatment of OCD, cognitive therapy procedures are sometimes used to identify faulty cognition, engage in cognitive restructuring and, with behavioral experiments, have the patient be exposed to situations that are designed to “disconfirm” the faulty cognition. Notably, exposure exercises provide the patient with the information that is needed to “correct” their distorted thinking without formal cognitive therapy.

Although CBT (specifically EX/RP) is recommended for children by Expert Concensus Guidelines and AACAP Practice Parameters (King et al., 1998; March, Frances, et al., 1997), there are no published large RCTs of CBT with children or adolescents. Uncontrolled CBT treatment trials reported to date suggest that most of the patients were responders, with a mean symptom decrease from 50% to 67%. March, Mulle, and Herbel (1994) reported on the outcome of a Type 3 study of a manualized and structured CBT protocol with 15 youth ages 8 to 18 years. There were significant differences between pre-and posttreatment, with a mean reduction in symptom severity of about 50%. Sixty-seven percent of the patients had better than 50% symptom improvement (which was maintained at follow-up). Only 20% were nonresponders. At the end of treatment, 40% were rated as asymptomatic, and at follow-up 60% were rated as such. Forty percent of the sample was able to discontinue medication with booster sessions. This work has led to a manual for children and adolescents with OCD (8 to 17 years) (March & Mulle, 1998), and an RCT comparing the relative efficacy of this CBT manual, sertraline, combined treatment, and placebo has just been completed (Franklin, Foa, & March, 2003); a second RCT comparing EX/RP plus family therapy to relaxation is also nearing completion (J. Piacentini, personal communication).

Franklin and colleagues (1998) conducted a Type 3 study and reported that 14 youth (ages 10 to 17 years) had a 67% improvement at posttreatment, and found no difference in response between those who received CBT alone and those with CBT in combination with an SRI. Others have also reported generally positive results for CBT in Type 3 open studies (Piacentini et al., 1994; Wever & Rey, 1997). Most of the studies used a weekly therapy regimen, generally over 3 months. Wever and Rey (1997) used a more intensive CBT protocol of 10 daily sessions over 2 weeks and found generally similar results. Franklin and colleagues (1998) reported that there was no difference in results, regardless of whether the treatment was delivered in 14 weekly sessions over 14 weeks or in 18 sessions over 4 weeks, although there was no random assignment.

In the only published direct Type 1 comparison of CBT and pharmacotherapy for pediatric OCD, de Haan, Hoogduin, Buitelaar, and Keijsers (1998) reported that whereas CBT and clomipramine both appeared to reduce OCD symptoms, CBT was superior to clomipramine. Until the systematic comparisons of single and combined treatments are completed, the relative efficacy cannot be directly compared, although the existing data suggest that CBT may provide more durability than medication.

The family context may play an important

role in treating OCD. Freeman, Garcia, and Leonard (2003) proposed that the family is a vehicle for treatment of OCD in youth and in the treatment of childhood anxiety disorders more broadly (Barrett, Dadds, et al., 1996; Howard & Kendall, 1996; Sanders, 1996; Siqueland & Diamond, 1998). Cognitive-behavioral therapy with family intervention can and often does involve parents and significant others as “behavior change agents.” Because the affective and cognitive aspects of the parent–child relationship, as well as the anxiety targets of intervention, are dealt with, the approach has been called cognitive-behavioral family therapy. Freeman and colleagues (2003) have manualized a treatment for children and their families with OCD, and initial pilot results are promising.

The systematic efficacy studies of the SRIs for the treatment of pediatric OCD form the largest body of work in the pharmacotherapy of the childhood psychiatric disorders, other than that of ADHD. An increasing literature supports the acute efficacy of clomipramine and the SSRIs in the treatment of children and adolescents with OCD.

The TCA clomipramine (an SRI) was the first medication systematically studied in children and adolescents with OCD. Three studies supported the efficacy of clomipramine for pediatric OCD (DeVeaugh-Geiss et al., 1992; Flament et al., 1985; Leonard et al., 1989). Flament and colleagues (1985) reported in a Type 1 study that in 23 youth in a 10-week double-blind, placebo-controlled, crossover study, clomipramine (in doses of 3 mg/kg) was significantly more effective than placebo in decreasing OCD symptomatology at week 5. For some patients, a reduction in symptoms was seen by 3 weeks. Of the 19 children who completed the trial, 75% had a moderate to marked improvement, and only 16% of the patients were unchanged. Extending this finding in an 8-week multicenter double-blind, parallel Type 1 study, DeVeaugh-Geiss and colleagues (1992) reported that clomipramine was superior to placebo for the treatment of OCD in youth. The 31 patients on clomipramine had a mean reduction in OCD severity scores of 37%, compared to 8% in the 29 patients on placebo. This study led to FDA approval for an SRI in pediatric OCD (children 10 years and older). The studies reported that clomipramine was generally well tolerated and had an anticholinergic adverse-effects profile. Periodic electrocardiograms (EKGs) are obtained during ongoing clinical care because of concerns about tachycardia and prolongation of the QTc interval.

To address whether the serotonergic specificity of clomipramine was critical, a double-blind crossover Type 1 study of clomipramine and desipramine (a selective noradrenergic reuptake inhibitor without serotonergic activity) was completed in 48 children and adolescents with OCD (Leonard et al., 1989). In this 12-week double-blind design, a patient received two weeks of placebo, then 5 weeks of clomipramine and 5 weeks of desipramine in a crossover. Clomipramine (targeting 3 mg/kg/day and not exceeding 5 mg/kg/day) was significantly better than desipramine in decreasing the OCD symptoms at week 5. Desipramine was no more effective in reducing OCD symptoms than placebo was in the Flament et al. (1985) study. The study noted that some of the patients who had received clomipramine in the first phase and desipramine in the second phase had a return of their symptoms when transferred to desipramine in the last phase, which suggests that the serotonergic reuptake inhibition may be critical.

With the development of the SSRIs, there is considerable discussion as to whether clomipramine is more effective than the SSRIs and at what point it should be chosen over the SSRIs. Initial meta-analysis (Griest, Jefferson, Kobak, & Katzelnick, 1995) indicated that clomipramine might be somewhat more effective than an SSRI for OCD in adults, but small head-to-head comparisons of clomipramine and fluvoxamine (Freeman, Trimble, Deakin, Stokes, & Ashford, 1994) and of fluvoxamine, paroxetine, and citalopram (Mundo, Bareggi, Pirola, & Bellodi, 1997) suggest that they have equivalent efficacy. The Expert Consensus Guidelines (March, Frances, et al., 1997) suggest that clomipramine be used when a patient has failed two or three adequate trials of an SSRI in combination with CBT.

The SSRIs have now emerged as the first-line

pharmacotherapeutic agent for OCD. They have the advantage over clomipramine of having a generally more tolerable side-effect profile with few anticholinergic effects, safer profile in overdoses, and no required EKG monitoring. Large multicenter Type 1 efficacy studies have shown that fluoxetine, fluvoxamine, and sertraline were each superior to placebo for children and adolescents with OCD (Geller et al., 2001; March, Biederman, et al., 1998; Riddle et al., 2001). Type 3 open studies support the use of citalopram and paroxetine (Rosenberg, Stewart, Fitzgerald, Tawile, & Carroll, 1999; Thomsen, 1997). The FDA has approved sertraline for the treatment of OCD in children ages 6 years and older, fluvoxamine for children ages 8 years and older, and fluoxetine for 8 years and older.

March, Biederman, and colleagues (1998) reported on 187 children and adolescents (ages 6 to 17 years of age) with OCD in a randomized double-blind, placebo-controlled, 8-week trial of sertraline (forced titration to 200 mg/day) versus placebo. Patients on sertraline, in comparison to those on placebo, showed significantly greater improvement on the several scales. Significant differences (with intent-to-treat analyses) between the two groups were seen as early as week 3 and continued for the entire study.

Riddle and colleagues (2001) reported the safety and efficacy of fluvoxamine for 120 youth (ages 8 to 17 years) with OCD in an RCT in which youth received either fluvoxamine (50–200 mg/day) or placebo for 10 weeks. Patients in the fluvoxamine group had a significant improvement (as measured on the Children's Yale-Brown Obsessive-Compulsive Scale, CY-BOCS) in comparison to the placebo group, and a difference could be measured as early as the first week. In the fluvoxamine group, 42% were responders (defined as a 25% decrease in measure of OCD symptom severity, CY-BOCS) in comparison to 26% in the placebo group, which was significantly different.

Geller and colleagues (2001) randomized 103 youth with OCD to either fluoxetine (starting at 10 mg/day) or placebo (in a 2:1 ratio) for 8 weeks. Intent-to-treat analyses reported that those in the fluoxetine group had significantly better improvement on CY-BOCs than did the placebo group. The authors concluded that fluoxetine at 20 to 60 mg daily was effective and well tolerated in the pediatric group. In contrast, Liebowitz and colleagues (2002) randomized 43 patients to either fluoxetine or placebo for 8 weeks. Responders then went into an 8-week maintenance phase. Fluoxetine dosage was fixed at 60 mg/day for 6 weeks and then could be increased to 80 mg/day. At week 8, fluoxetine was not significantly better than placebo on the CY-BOCs or CGI-Improvement scale; authors attributed this to either low power or short duration of treatment. The fluoxetine group continued to improve during the maintenance phase such that at week 16, 57% of the fluoxetine patients, as compared to 27% of the placebo patients (using data at week 8), were much or very much improved. The authors concluded that fluoxetine's effect took more than 8 weeks to develop.

Review of adverse effects of the SSRIs suggests that dropouts from blinded active medication assignment are usually less than 13%, and in many studies there are no significant differences between the dropouts receiving medication or placebo (March, 1999). Generally, the most common side effects seen with the SSRIs include sedation, nausea, diarrhea, insomnia, anorexia, tremor, sexual dysfunction, and hyperstimulation (March, Biederman, et al., 1998; Riddle et al., 2001). Children and adolescents may be more vulnerable to agitation or activation while on SSRIs than are adults, but this is not well studied. Rare adverse reactions include apathy syndrome, serotonin syndrome, and extrapyramidal symptoms. Pharmacokinetic studies of sertraline (Alderman, Wolkow, Chung, & Johnston, 1998) and of paroxetine (Findling et al., 1999) reported wide intra-and between-individual pharmacokinetic variability but generally similar results to those reported in adults.

How large is the treatment response on an SSRI? The pediatric treatment response is similar to that reported in adults. In general, a 30%–40% reduction in OCD symptoms, which corresponds to an average 6-to 8-point decrease on the CY-BOCS (Scahill et al., 1997), is reported in the medication treatment group in the SSRI controlled studies (March, 1999). Unlike some of the other disorders, there is little or no placebo effect reported. Clinical benefits may begin as early as 3 weeks and typically plateau at about 10 weeks (March, 1999).

The Expert Consensus Guidelines (March, Frances, et al., 1997) and the AACAP Practice Parameters for OCD (King et al., 1998) both recommend starting treatment in children with CBT or CBT plus an SSRI, depending on severity and comorbidity. Both guidelines recommend that patients started on SSRI alone who are partial responders should have a trial of CBT. The study of the relative efficacy of medication, of CBT, and their combination has just been completed using a Type 1 randomized controlled trial (Franklin et al., 2003; Pediatric OCD Collaborative Study Group, 2004; Foa & March, personal observations).

Hypotheses concerning whether Sydenham's chorea and PANDAS might share similarities in their pathophysiology have led to the question of whether penicillin prophylaxis would reduce neuropsychiatric symptom exacerbation in children with PANDAS by preventing streptococcal infection. An 8-month, Type 1 double-blind, placebo-controlled, crossover trial of oral penicillin V (250 mg twice daily) and placebo was conducted in 37 children (Garvey et al., 1999). There was no significant difference between phases in either the OCD or tic symptom severity; however, penicillin administration failed to provide adequate prophylaxis against GABHS (as evidenced by the fact that 14 of 35 GABHS infections occurred during the penicillin phase). A number of children received antibiotic treatment multiple times during the placebo phase. The authors concluded that because of the failure to achieve an acceptable level of streptococcal prophylaxis, no conclusions could be drawn regarding the efficacy of penicillin prophylaxis in preventing tic or OCD symptom exacerbation. Further studies are needed that use a more effective prophylactic agent and include a larger sample size. Clinical experience would recommend workup for GABHS infection in children with abrupt and sudden onset of OCD or tics and dramatic exacerbations (Murphy & Pichichero, 2002; Swedo et al., 1998).

If post-streptococcal autoimmunity is the cause of the exacerbations in this subgroup, then children with PANDAS might benefit from im munomodulatory therapies that have been shown in preliminary findings to treat symptoms of Sydenham's chorea. In a Type 1 study, children with severe, infection triggered exacerbations of OCD or tic disorders were randomly assigned to plasma exchange (five single-volume exchanges over 2 weeks), intravenous immunoglobulin (IVIG;1 g/kg daily on 2 consecutive days), or placebo (saline solution given in the same manner as IVIG). Plasma exchange and IVIG were both effective in lessening of symptom severity for this group of children. Ratings were completed at 1 month and symptom gains were maintained at 1 year (Perlmutter et al., 1999). It should be noted that these children were more much significantly impaired than the average child with OCD or tics, and thus these invasive interventions were considered. These interventions are investigational and should only be considered in the context of research approved by a human investigations committee and not in the context of routine clinical care (Leonard & Swedo, 2001).

With respect to psychotherapy for addressing posttraumatic stress disorder (PTSD), treatments derived from behavioral and cognitive theoretical models have been most carefully examined. In RCTs, the efficacy of several CBT treatments has been evaluated, including prolonged exposure (PE), anxiety management, eye movement desensitization and reprocessing (EMDR), cognitive therapy, and treatment packages that combine approaches.

Summaries of the studies of PE with veterans (e.g., Cooper & Clum, 1989; Glynn et al., 1999; Keane, Fairbank, Caddell, & Zimmering, 1989a) have consistently indicated that although exposure therapy is superior to various control conditions, the magnitude of the improvement has been somewhat limited. Because of the extent and chronicity of their problems and the presence of incentives for the veterans to minimize acknowledging treatment gains (e.g., losing service-connected disability compensation),