A retrospective medical chart review of clinical outcomes in children and adolescents with attention-deficit/hyperactivity disorder treated with guanfacine extended-release in routine Canadian clinical practice

This study was a retrospective medical chart review in children and adolescents with ADHD recruited from 10 sites by Canadian community- or hospital-based pediatricians, child psychiatrists, or family doctors specializing in ADHD. Included children (6–12 years) and adolescents (13–17 years) had diagnosed ADHD, had initiated treatment with GXR (monotherapy or adjunctive therapy), and had ≥ 6 months of medical chart data post-initiation of GXR (with ≥ 1 follow-up; if applicable, this included the visit when GXR was discontinued). Patients could have been treatment naive or received other pharmacotherapy before GXR initiation. In treatment-naive patients, the observation period of the chart review began at treatment initiation, whereas in patients who had received prior ADHD pharmacotherapy, the observation period started 12 months prior to GXR initiation. The chart review observation period ended at ≤ 12 months after GXR initiation. If GXR was discontinued, the reason for discontinuation was noted but no other follow-up data were extracted after the discontinuation visit.

Data were transcribed from patient charts to case report forms (CRFs) by the site investigator or designated staff and sent to a central database. No patient identifying information was recorded; patients were given encrypted random study identification numbers. Data on patient demographics and clinical characteristics were extracted from patient charts per standard of care for ADHD. At ADHD diagnosis, these data included diagnosis date, sociodemographics (age, gender, ethnicity, location of residence [urban vs rural]), medical history, and comorbidities. At GXR initiation, data extracted included reason for GXR initiation, prior non-pharmacologic treatments for ADHD or comorbid conditions, and prior pharmacologic treatments, including stimulants, non-stimulants, and atypical antipsychotics (AAPs) for ADHD or comorbid psychiatric conditions/symptoms. At GXR initiation and for its duration, covariates included GXR information (dose, frequency, dose/frequency changes and reasons, termination date and reason) and pharmacologic treatment for ADHD or comorbid psychiatric symptoms. If GXR was discontinued, reasons for discontinuation and any subsequent ADHD-related pharmacologic treatment regimen were recorded, and no additional follow-up details were extracted following the discontinuation visit.

Extracted outcome data were from the time prior to (for patients receiving prior pharmacologic treatment only), at GXR initiation, and for its duration (≤ 12 months). ADHD symptoms and functionality were measured using the General Physician Symptom Assessment and General Physician Functional Assessment (school grades/performance and homelife), respectively. Rating scale data on ADHD symptoms and functionality, if available in patient charts, were also extracted and evaluated. Symptom ratings scales included Swanson, Nolan, and Pelham-IV (SNAP-IV-26) [27], National Institute for Children’s Health Quality (NICHQ) Vanderbilt Assessment Scale and Follow-up-Parent Information [28], and Conners 3-Parents (Conners 3-P) [29]. The Weiss Functional Impairment Rating Scale-Parent Report (WFIRS-P) was used to measure functionality [30]. To minimize recall bias related to subjective assessment of symptomatology (as extracted from patients’ charts in the absence of validated scales), outcome recorders were requested to categorize consistently the patient’s response, as indicated in the patient charts. Outcome recorders had prior training including how to record consistently a category for patient responses based on data in patient charts. ADHD symptoms and functionality were categorized as “improvement”, “no change”, or “worsening” based on rating scales in medical charts, and/or clinician’s notes. An improvement/worsening of symptoms or functionality for two consecutive visits was categorized as “improved” or “worsening”, respectively.

Treatment-emergent adverse events (TEAEs) occurring during GXR treatment were coded using the Medical Dictionary for Regulatory Activities (MedDRA) version 20 and described by the proportion of patients experiencing a TEAE within each Preferred Term and System Organ Class.

Charts from a minimum of 252 and maximum of 444 patients were required. In research by Cutler et al. [31], symptomatic remission was achieved in 62.2% versus 46.1% of patients with ADHD aged 6–17 years with suboptimal response to stimulants treated with GXR plus stimulant versus placebo plus stimulant, respectively. Based on these data, the assumption was that ~ 60% of patients in this study would experience symptom improvement. To detect a significant difference in response with α = 5% and 80% power, 202 evaluable patients were required. Allowing that required data would not be available in 25% of charts, data from a minimum of 252 charts were required.

In Cutler et al. [31], 77.3% of patients experienced a TEAE. Assuming a similar rate in our study with 252 patients, the precision of estimate as assessed with the width of the 95% confidence interval (CI) would be 13.5% of the estimate (0.104), which is within acceptable limits. To achieve 10% precision (0.078), 444 patients would be required. With 252 and 444 patients, the probability of detecting a TEAE with true incidence of 1% was 92% and 99%, respectively.

Summary statistics were calculated for all study variables, including mean, median, standard deviation (SD), 95% CI of the mean, and frequency distributions for continuous scale variables.

The primary analysis was change in ADHD symptoms and functionality, assessed by the proportion of patients showing improvement versus no change versus worsening. For patients without previous ADHD treatment, this categorization was based on assessment of the patient compared with the previous visit (beginning with the comparison of the first assessment post-GXR initiation to status at initiation). For patients who had received prior ADHD pharmacologic treatment, the pre-GXR value corresponded to the overall response (improved, no change, or worse) to the last treatment regimen administered prior to GXR initiation. An improvement/worsening of symptoms or function was defined as a categorization of patient response as improved or worse, respectively, for two consecutive visits. The best symptom and functionality responses were described, corresponding to the best respective responses registered during the post-GXR initiation observation period while receiving GXR.

Post-hoc analyses (unspecified in the protocol) included assessments of symptoms and functionality whether GXR treatment was monotherapy/adjunctive therapy, if patients had been initiated on GXR to reduce use of AAPs, and according to psychiatric comorbidities of interest (oppositional defiant disorder, learning disability, anxiety, and autism spectrum disorder).

Data from patients who had received ADHD pharmacologic treatment before GXR and had ≤ 12 months of data following GXR initiation, a subset not considered in power calculations, were examined in exploratory analyses. The objective was to assess the impact on clinical outcomes of switching to GXR or adding GXR to current ADHD treatment. The pre-GXR value (prior period) corresponded to the overall response (improved, no change, or worse) to the last treatment regimen administered before GXR initiation. The analyses detailed for the primary objective were repeated in this subgroup, although the post-hoc breakdown of results (by monotherapy/adjunctive therapy, etc.) was not conducted. The comparison of the response rate (improvement vs no change vs worsening) of General Physician Symptom and Functional Assessments before and after GXR initiation was conducted with the extension of McNemar’s test (Bowker’s test); two-tailed test. The null hypothesis was that the percentages are equal before and after GXR initiation, i.e., during the baseline period and post-GXR initiation observation period (exploratory analysis).

As this was a retrospective chart review, no missing data could be retrieved. However, cross-validation was performed, and erroneous data were clarified with physicians. Partial dates were completed using the most conservative approach when required to determine duration of observation, treatment exposure, and incidence of adverse events. Otherwise, all analyses were conducted on available data with no data imputation.

In total, 330 charts were screened, and all patients were included in the analysis (Additional file 1: Table S1). Approximately 92% and 40% had prior pharmacologic and non-pharmacologic ADHD treatment within the 12 months preceding GXR initiation, respectively. The majority of patients were male Caucasian children (6–12 years) living in Ontario.

Comorbid psychiatric conditions were reported for 215/330 (65.2%) patients. The most frequently reported conditions experienced by ≥ 5% of patients were oppositional defiant disorder (n = 92, 27.9%), learning disability (n = 70, 21.2%), anxiety (n = 53, 16.1%), autism spectrum disorder (n = 35, 10.6%), and tic (n = 18, 5.5%).

A total of 270 (81.8%) patients concomitantly used other pharmacological treatments for ADHD and comorbid psychiatric conditions while on GXR. The most frequently used medications included methylphenidate (n = 123, 37.3%), lisdexamfetamine (n = 121, 36.7%), melatonin (n = 26, 7.9%), centrally acting sympathomimetics (n = 24, 7.3%), apriprazole (n = 25, 7.6%), and risperidone (n = 18, 5.5%). Six (1.8%) patients used clonidine, 1 (0.3%) patient used trazodone, and 40 patients used SSRIs (citalopram, n = 3, 0.9%; escitalopram, n = 13, 13.9%; fluoxetine, n = 16, 4.8%; paroxetine, n = 1, 0.3%; sertraline, n = 7, 2.1%).

The most common reasons for GXR initiation were non-optimal control of ADHD symptoms, to improve response to stimulants, to extend duration of effect of stimulants, and to avoid increasing doses of ADHD stimulants (Additional file 1: Table S2). Compared with 60 (18.2%) patients receiving monotherapy, 270 (81.8%) received GXR adjunctive therapy. The median daily dose of GXR at initiation was 1.0 mg. Of the 325/330 (98.5%) patients who changed GXR dose during post-GXR initiation observation, 307/325 (94.5%) had an increase and 18/325 (5.5%) had a decrease. The mean (SD) total dose change for all patients was 2.4 mg (0.5 mg). The mean (SD) length of GXR treatment was 7.9 (3.6) months and was similar between children (n = 242) and adolescents (n = 88). After a median (range) of 8.4 (0.2–12) months, almost 80% of patients were continuing GXR treatment (~ 20% discontinued). The most common reasons for discontinuation were safety/tolerability issues.

Improvements in clinical outcomes were reported for > 70% of patients receiving GXR (Fig. 1). No change/worsening of ADHD symptoms was observed for < 5% of patients receiving GXR (missing data, 77 [23.3%]). Additionally, > 60% of patients showed improvements in school performance and homelife while no change/worsening in school performance and homelife functionality was reported for < 10% of patients.

Overall, 349 TEAEs were experienced by 147/330 (44.5%) patients (Table 1). The majority were non-serious (341/349, 97.7%), mild in severity (n = 314, 90.0%), and related to GXR treatment (n = 263, 75.4%). No action was deemed necessary regarding management of GXR treatment for 53% of TEAEs. The most common TEAEs (reported in ≥ 5% of patients) were somnolence (n = 59/330, 17.9%), headache (n = 34/330, 10.3%), insomnia (n = 26/330, 7.9%), presyncope (n = 22/330, 6.7%), and decreased appetite (n = 18/330, 5.5%). Eight serious TEAEs were experienced by 7 patients (Table 1). Of these, 6/8 events were moderate in severity, 7/8 were related to GXR treatment, and 7/8 were reported as resolved. A total of 4 serious TEAEs experienced by 3 patients led to GXR discontinuation, 1 serious TEAE resulted in dose interruption, and 1 serious TEAE resulted in dose reduction. No action was deemed necessary regarding GXR management for 2 serious TEAEs experienced by 2 patients.

Improvements in ADHD symptoms were observed in 35/60 (58.3%) patients receiving monotherapy and 197/270 (73.0%) patients receiving GXR adjunctive therapy. Similar improvements were seen in functionality assessments of school performance (monotherapy, 30/60 [50.0%]; adjunctive therapy, 183/270 [67.8%]) and homelife (monotherapy, 30/60 [50.0%]; adjunctive therapy, 179/270 [66.3%]).

Of the 44 patients initiated on GXR to reduce use of AAPs, ADHD symptom improvements were experienced by > 70%, including > 55% of patients in school performance and homelife functionality (Fig. 2). AAPs used included risperidone, quetiapine, and aripiprazole.

Patients with comorbid autism spectrum disorder, oppositional defiant disorder, anxiety, or learning disability experienced improvements in ADHD symptoms and functionality consistent with the primary findings (Table 2).

In the subgroup who had received prior pharmacotherapy, a greater proportion of patients who had reported worsening ADHD symptoms on prior pharmacotherapy experienced improvements on GXR compared with those reporting no changes (Table 3). Similarly, a greater proportion of patients reporting worsening changes in homelife functionality and school performance on prior pharmacotherapy experienced improvements on GXR compared with those reporting no changes.