This open trial conducted by the Research Units on Pediatric Psychopharmacology (RUPP) Autism Network showed solid reductions in parent- and teacher reported hyperactive-impulsive behaviors in children unresponsive to previous trials of a stimulant. Symptom decreases from baseline were 25–36% by 8 weeks, depending on the source of ratings, and 49% of subjects were rated by study clinicians as clinically significantly improved on global measures (Scahill et al. 2006).

Despite the apparent promise of guanfacine for the treatment of hyperactive-impulsive behaviors in this population, prominent variability in treatment efficacy has been noted. In the Scahill et al. report (2006), the improvement from baseline in Swanson, Nolan, and Pelham (SNAP) total ADHD symptom scores ranged from −1 to 41 points (mean 12.8). Similarly, in a large chart review study of children with PDD and other disorders receiving guanfacine, the rate of positive response ranged from 13% to 39% (average 23%), depending on the patient population (Posey et al. 2004). In a sample of ADHD children with co-occurring tic disorders, guanfacine was superior to placebo and reduced ADHD symptoms by 37% from baseline, but end-point ADHD ratings also showed large interindividual variability (Scahill et al. 2001). Another small placebo-controlled trial in children with tic disorders also showed evidence of interindividual variability in response (Cummings et al. 2003)

Following a detailed screening that included medical, psychiatric, and developmental assessments, eligible subjects were seen weekly for the first 4 weeks to evaluate response and tolerability and then every other week until week 8. Visits included: Vital signs, height and weight, and a systematic review of adverse events. Outcome ratings were collected at baseline, and weeks 2, 4, 6, and 8 of treatment.

Overall parent ratings showed declines in hyperactive-impulsive symptoms during the 8 weeks (see Table 1). Parent-rated Hyperactivity on the ABC–Hyperactivity scale declined 58% from 29.5±8.5 at baseline to 17.3±9.4 at end point. On the parent-rated SNAP, mean scores declined 56% from 37.0±7.8 at baseline to 21.8±9.5 at end point.

Because of prior data showing most pronounced effects of gene variants on gene expression for the T/T homozygotic condition (Kimchi-Sarfaty et al. 2007), we contrasted the change across baseline to 8 weeks using all outcome points for hyperactive-impulsive symptoms for two groups, the C/C and C/T versus the T/T subjects. Allele frequencies were 55% for the C and 45% for the T allele, respectively; the allele frequencies observed were comparable to other reports (Maeda and Sugiyama 2008). Genotypes were grouped as C/C or T/C (n=16) versus T/T (n=6). MDR1 genotype frequencies were in Hardy–Weinberg equilibrium; χ²=2.72, p=0.10. Genotype groups did not differ in study completion rates, age, race, gender, final dose, or SNAP or ABC–Hyperactivity baseline symptom ratings (see Table 2), but the T/T genotype showed a significantly smaller drop in the SNAP and ABC–Hyperactivity summary score versus the C/C and C/T subjects (mean decrease of 22.5% and 16.2% versus 47.9% and 50.0%;),; SNAP summary and ABC–Hyperactivity, respectively) (Fig. 1). The amount of variation in drop of ABC–Hyperactivity and SNAP summary scores over 8 weeks that can be attributed to the MDR1 genotype was 53% and 40%, respectively. Effect sizes for guanfacine on SNAP total scores calculated from pre- versus posttreatment differences divided by average standard deviations (SDs) for the C/C and C/T group was 1.73 versus 0.54 for the T/T subjects. Similarly, effect sizes for guanfacine on ABC parent-rated Hyperactivity were 1.48 versus 0.31 for C/C and C/T versus T/T groups, respectively.

MDR1 is a member of the ABC family of membrane-bound drug transporter proteins that are present at the blood–brain barrier. Besides MDR1 (ABCB1), the group includes multidrug resistance (-associated) protein 1 (MRP1, ABCC1), MRP2 (ABCC2), MRP3 (ABCC3), MRP4 (ABCC4), MRP5 (ABCC5), and the breast cancer resistance protein (BRCP, ABCG2) (Wang et al. 2007). The substrates and inhibitors of each of these transporters vary widely, but psychotropic drugs have been found to bind to many of these proteins. In the case of the best-studied member of the family, MDR1, 29 SNPs have been identified. The MDR1 variant we chose to examine has previously been shown to influence gene expression (Hoffmeyer et al. 2000) and substrate specificity (Kimchi-Sarfaty et al. 2007). Indeed, one report described a reduction of nearly 50% in MDR1 protein concentration between the C/C versus the T/T genotype; however, not all reports have observed effects on expression (Sakaeda 2005). The effects of such reductions in MDR1 can be complex. Reduced MDR1 has been associated with increased brain concentrations of drugs known to be transported by MDR1; however, reduced MDR1 can also reduce absorption and increase renal excretion of some substrates, defying straightforward attempts to explain differences in clinical effects by genotype.

Our report should be viewed as exploratory, given the small sample composed only of individuals previously shown to be unresponsive to or intolerant to MPH, with a restricted age range, predominance of males and Caucasian subjects, a lack of treatment blinding, and short observation period. Data from much larger clinical trial samples are needed to confirm or refute this observation and to better estimate the magnitude of the difference, if any, associated with the MDR1 genotype. Another limitation is the absence at present of definitive in vitro data confirming that guanfacine is a major substrate for MDR1 transport. If confirmed that guanfacine transport is influenced by MDR1, it would also be critical to examine in a controlled system the precise impact of MDR1 gene variants on guanfacine absorption, distribution, blood–brain levels, and excretion. Our ability to examine MDR1 genotype effects on adverse events is also limited in this pilot study. However, a large clinical trial sample could examine whether more commonly observed adverse events, such as sedation or mood liability, adverse events that we observed often mandated dose reduction or predicted intolerance, may also be associated with the MDR1 genotype. Such observations hold obvious potential clinical importance if applicable at the level of the individual patient.

Clinically, additional trials are needed to determine the benefits and safety of the α agonists as common treatments for pediatric neuropsychiatric disorders, especially in individuals with PDD with ADHD symptoms. More data are needed on the durability of benefits, as some reports suggest waning effects over time (Jaselskis et al. 1992), effects of sedation and fatigue on adherence (Biederman et al. 2008b), and to confirm the apparent acceptable cardiovascular safety profile from larger, recent reports (Biederman et al. 2008a; Daviss et al. 2008). As clinical support for guanfacine is growing, given the modest benefits of stimulants in subgroups of children such as PDD (Aman et al. 2003; Research Units on Pediatric Psychopharmacology [RUPP] Autism Network 2005), it is important to attempt to identify possible predictors of treatment benefit. Accumulated findings from initial guanfacine trials in children for the management of ADHD behaviors have varied, perhaps due to differences in clinical populations sampled. Although our data support MDR1 gene variants as significant moderators to guanfacine response, there yet may be other important variants in other genes that further contribute to variability in guanfacine response. Examination of other relevant target genes thought to play a role in guanfacine distribution, neurochemical effect, or metabolism, including α2A (ADRA2A), other noradrenergic system genes, CYP450 genes, and others could well reveal more robust genetic influences on clinical response to guanfacine.