Predictors of treatment outcome in adults with ADHD treated with OROS® methylphenidate

Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

Contents lists available at ScienceDirect

Progress in Neuro-Psychopharmacology & Biological Psychiatry j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p n p

Predictors of treatment outcome in adults with ADHD treated with OROS® methylphenidate Jan K. Buitelaar a,⁎, J.J. Sandra Kooij b, J. Antoni Ramos-Quiroga c, Joachim Dejonckheere d, Miguel Casas e, Joop C. van Oene f, Barbara Schäuble g, Goetz-Erik Trott h a

UMC St. Radboud, Department of Cognitive Neuroscience, Nijmegen, The Netherlands PsyQ Psycho-Medical Programs, Program Adult ADHD, Den Haag, The Netherlands Universitat Autònoma de Barcelona, Hospital Universitari Vall d'Hebron, Department of Psychiatry, Adult ADHD Program, Barcelona, Spain d SGS, Mechelen, Belgium e Hospital Universitari Vall d'Hebron, Servicio de Psiquiatría, Barcelona, Spain f Janssen-Cilag EMEA Medical Affairs, Tilburg, The Netherlands g Janssen-Cilag EMEA, Psychiatry, Neuss, Germany h Private Practice Aschaffenburg, Germany b c

a r t i c l e

i n f o

Article history: Received 5 October 2010 Received in revised form 10 December 2010 Accepted 16 December 2010 Available online 23 December 2010 Keywords: Attention deficit hyperactivity disorder Adults OROS-methylphenidate Response

a b s t r a c t Background: We conducted a post-hoc analysis of the Long-Acting MethylpheniDate in Adult attention-deficit hyperactivity disorder (LAMDA) study to investigate predictors of response in adults with ADHD randomly assigned to Osmotic Release Oral System (OROS)®-methylphenidate hydrochloride (MPH) 18, 36 or 72 mg or placebo. Methods: LAMDA comprised a 5-week, double-blind (DB) period, followed by a 7-week, open-label (OL) period. A post-hoc analysis of covariance and a logistic regression analysis were undertaken to detect whether specific baseline parameters or overall treatment compliance during the double-blind phase contributed to response. The initial model included all covariates as independent variables; a backward stepwise selection method was used, with stay criteria of p b 0.10. Six outcomes were considered: change from baseline CAARS: O-SV (physician-rated) and CAARS:S-S (self-report) scores at DB and OL end points, and response rate (≥ 30% decrease in CAARS:O-SV score from baseline) and normalization of CAARS:O-SV score at DB end point. Results: Taking into account a significant effect of OROS®-MPH treatment versus placebo in the original analysis (p ≤ 0.015), across the outcomes considered in this post-hoc analysis, higher baseline CAARS scores were most strongly predictive of superior outcomes. Male gender and lower academic achievement were also predictive for improved results with certain outcomes. Conclusions: Several baseline factors may help to predict better treatment outcomes in adults receiving OROS®-MPH; however, further research is required to confirm these findings and examine their neurobiological underpinnings. © 2010 Published by Elsevier Inc.

1. Introduction

Abbreviations: ADHD, attention-deficit hyperactivity disorder; ANCOVA, analysis of covariance; CAADID, Conners' Adult ADHD Diagnostic Interview for DSM-IV; CAARS:OSV, Conners' Adult ADHD Rating Scale (physician rated); CAARS:O-S-S, Conners' Adult ADHD Rating Scale (self-reported, short version); CI, confidence interval; DB, doubleblind; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth edition; LAMDA, Long-Acting MethylpheniDate in Adult attention-deficit hyperactivity disorder; LOCF, last observation carried forward; MPH, methylphenidate hydrochloride; NICE, National Institute for Health and Clinical Excellence; OL, open-label; OR, odds ratio; OROS, Osmotic Release Oral System; SE, standard error. ⁎ Corresponding author. Donders Institute for Brain, Behavior and Cognition, Radboud University, Nijmegen Medical Center, Department of Cognitive Neurosciences (204), PO Box 9101, 6500 HB Nijmegen, The Netherlands. Tel.: + 31 24 3610750; fax: + 31 24 3610989. E-mail address: [email protected] (J.K. Buitelaar). 0278-5846/$ – see front matter © 2010 Published by Elsevier Inc. doi:10.1016/j.pnpbp.2010.12.016

Attention-deficit hyperactivity disorder (ADHD) is a heterogeneous and highly heritable disorder, manifesting itself in symptoms of inattention and/or hyperactivity/impulsivity that arise during childhood, frequently persistent throughout development into adulthood, and result in impairment in multiple domains of functioning (Barkley et al. 2002; Faraone et al. 2005; Biederman et al. 2006a; Faraone et al. 2006; Fried et al. 2006; Li et al. 2006). Pharmacotherapy is central in the management of ADHD, with methylphenidate hydrochloride (MPH) a cornerstone of treatment. MPH is recommended as first-line therapy in the treatment of children, adolescents and adults with ADHD based on a recent evaluation made by the British National Institute for Health and Clinical Excellence (NICE) (National Institute for Health and Clinical Excellence 2008). Notably, results from the

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

LAMDA study (Long-Acting MethylpheniDate in Adult attentiondeficit hyperactivity disorder) [protocol 42603ATT3002], which was a large, 5-week, double-blind, placebo-controlled study (Medori et al. 2008), with a 7-week, open-label extension phase (Buitelaar et al. 2009), demonstrated the efficacy, safety and tolerability of OROS®MPH in adults with ADHD. In this study, significant improvements in the scores in both the physician-rated version of Conners' Adult ADHD Rating Scale (CAARS:O-SV) and the self-reported, short version (CAARS:S-S) from baseline to double-blind end point were observed for all OROS®-MPH dosages (18 mg, 36 mg or 72 mg) compared with placebo (p ≤ 0.015 for all comparisons versus placebo); as expected, results for physician-rated and self-rated CAARS scores showed similar trends. Likewise, the proportion of responders (≥ 30% reduction in CAARS:O-SV total score) was significantly higher in the OROS®-MPH groups compared with placebo (Medori et al. 2008). In addition, the safety and tolerability profile in adults was comparable to that observed in children and adolescents. Although generally considered effective, reported response rates to MPH in adults with ADHD are quite variable (Kooij et al. 2004; Rösler et al. 2004; Kessler et al. 2006; Rösler et al. 2009). The reasons behind variability in response to therapy in ADHD are not fully understood. Associated co-morbidities have been shown to alter response to ADHD therapy (Ghuman et al. 2007; Newcorn 2009), while other factors, such as female gender, higher IQ, considerable inattentiveness, younger age, lower disease burden and compliance with stimulant medication may also account for variability of treatment response in children and adolescents with ADHD (August et al. 1983, Buitelaar et al. 1995; Hechtman 1999; van der Oord et al. 2008, MTA Cooperative Group, 1999). It has, however, been consistently observed that higher dosing of stimulants result in a higher percentage of responders (Spencer et al. 1995; Faraone et al. 2004; Biederman et al. 2006b; Medori et al. 2008). Whilst results from the LAMDA study demonstrate that adults with ADHD generally show good symptomatic improvement with MPH therapy, predicting response in individual patients above and beyond medication status remains somewhat elusive. Therefore, to further knowledge in this area, we report the results of a post-hoc analysis of the aforementioned LAMDA study, which was undertaken to investigate the influence of baseline characteristics and treatment variables on clinical outcomes in adults with ADHD. 2. Methods 2.1. Study design and patients This study has been previously described in detail. In brief, this was a 5-week, double-blind, randomized, placebo-controlled, parallelgroup, 4-arm, fixed-dose trial (Medori et al. 2008), followed by a 7week, open-label phase (Buitelaar et al. 2009) in adult men and women (aged 18–65 years) with a diagnosis of ADHD according to Diagnostic and Statistical Manual of Mental Disorders, Fourth edition (DSM-IV), criteria. Presence and chronicity of childhood symptoms were confirmed by the Conners' Adult ADHD Diagnostic Interview for DSM-IV (CAADID) (Conners C et al. 1999). Other requirements for inclusion were a CAARS:O-SV score of ≥24 points at screening. Following screening and a washout period of up to 4 weeks during which current therapy was tapered to discontinuation, eligible patients were randomly assigned into one of four treatment groups to receive once-daily oral dosages of 18 mg, 36 mg or 72 mg OROS®MPH or placebo. Patients in the 72 mg OROS®-MPH group were titrated from a starting dose of 36 mg/day for 4 days to 54 mg/day for 3 days, after which 72 mg/day was administered for 4 weeks. Following the double-blind treatment phase, eligible patients entered the 7-week, flexible-dose, open-label extension phase with OROS®-MPH, during which all patients were initiated on OROS®-MPH at 36 mg/day, with the exception of the German patients who started

555

with 18 mg/day. Patients were flexibly dosed, based on investigator assessment, between 18 and 90 mg/day total daily dose. Participants gave written informed consent. The study was reviewed and approved by the institutional review board of each participating centre. All procedures were conducted in accordance with the ethical standards of the responsible committee on human experimentation and the Declaration of Helsinki, 1983 (see for the latest version www. wma.net). 2.2. Statistical analyses 2.2.1. Predictive value of baseline characteristics A post-hoc analysis of covariance (ANCOVA) and a logistic regression analysis were undertaken to detect which baseline parameters contributed to the effect of OROS®-MPH or, in other words, which baseline parameters could be considered as predictors of treatment response. Six different outcome measures were considered: change from baseline in CAARS:O-SV score at the end of the 5-week, double-blind treatment period (1), or at the end of the subsequent 7-week, open-label extension phase (2); change from baseline in CAARS:S-S score: short version score at the end of the 5week, double-blind treatment period (3), or at the end of the subsequent 7-week, open-label extension phase (4); response rate at the end of the 5-week, double-blind treatment period, where response was defined as a 30% or greater decrease in CAARS:O-SV score from baseline (5); and rate of normalization of CAARS:O-SV score at the end of the 5-week, double-blind treatment period, where normalization was defined as a return of CAARS:O-SV score to within the normal range according to the specifications provided in the CAARS manual (6). A last observation carried forward (LOCF) imputation method was utilized for missing values. The CAARS:O-SV comprises 18 investigator-rated items corresponding to the 18 DSM-IV ADHD symptoms and provides a total score referred to as the CAARS total ADHD symptom score and two subscales. As patient ratings are a valuable source of additional data, the CAARS:S-S was also utilized. This is a 26-item, self-reported, four-point rating scale that measures symptoms based on the DSM-IV criteria for ADHD, providing a total score, ADHD index and four subscales. Change in CAARS:O-SV and CAARS:S-S scores (absolute measure) were assessed as continuous efficacy parameters using ANCOVA models; response (based on relative change) and normalization rates were assessed as binary efficacy parameters using logistic regression. The following nine parameters, which were assessed at screening or baseline, were used in the analysis: (1) age (years), (2) gender (male/female), (3) history of mood or anxiety disorders (yes/no), (4) history of drug or alcohol abuse (yes/no), (5) country (Czech Republic, Denmark, Finland, France, Germany, Greece, The Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom), (6) highest education level (on a 4-point scale: primary school, secondary school, high school or university); (7) employment status at baseline (yes/no); (8) baseline score of CAARS:O-SV or CAARS:S-S, respectively; and (9) randomization treatment group (placebo, OROS®-MPH 18 mg, OROS®-MPH 36 mg and OROS®-MPH 72 mg). Overall treatment compliance during the double-blind phase was also included in the analysis as the only non-baseline variable; compliance was calculated as the number of tablets actually taken during the double-blind phase, divided by the scheduled number of tablets. Baseline age and CAARS:O-SV and CAARS:S-S scores, and overall treatment compliance were included as continuous variables, with the other parameters analysed as categorical variables. The relation between the predictors and each of the outcomes was evaluated using correlation coefficients or visual inspection of cross-tabulations as applicable. All analyses used a stepwise approach. The initial model included all parameters as independent variables. The least significant covariate was then eliminated and the analysis repeated with one

556

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

less parameter. This process was repeated until all predictors remaining in the model had a significant contribution at the level of p value b0.10. Randomization group and baseline scores (CAARS:O-SV or CAARS:S-S) were retained in the model, regardless of whether or not their contribution was statistically significant. Variables surviving these stepwise analyses are discussed below. 3. Results 3.1. Patient characteristics and disposition A total of 448 patients were screened and 401 patients were randomly assigned and treated in the double-blind phase of the study (Medori et al. 2008). Overall, 365 (91%) randomized patients completed the 5-week, double-blind study period. Baseline demographic and clinical characteristics were similar across placebo and the three OROS®-MPH groups (Table 1). A total of 370 patients entered the open-label treatment phase and 337 (91%) completed it (Buitelaar et al. 2009). Demographic and baseline characteristics of the open-label population were similar to those of the double-blind population. During the open-label extension, the median final daily dose of OROS®-MPH was 54 mg. 3.2. Efficacy and safety during the overall study Full efficacy and safety results have been reported previously (Medori et al. 2008; Buitelaar et al. 2009). During the double-blind phase, treatment with OROS®-MPH was associated with significantly larger improvements in CAARS:O-SV and CAARS:S-S total symptom score from baseline to end point (LOCF) than placebo (p ≤ 0.015 for all), with improvements continuing in the open-label phase (Buitelaar et al. 2009). OROS®-MPH was well tolerated throughout both phases, with few patients discontinuing due to adverse events.

3.3. Predictors of treatment outcome 3.3.1. Conners' Adult ADHD Rating Scale (CAARS:O-SV) 3.3.1.1. Double-blind phase. In this post-hoc analysis (Table 2), increasing age was associated with greater improvements in CAARS:O-SV scores (ANCOVA regression coefficient [β] [standard error (SE)]= −0.105 [0.050]; p = 0.0358). A difference in response to PR-OROS®MPH was also observed between the genders: male patients were more likely than female patients to experience improvements in CAARS:O-SV scores at double-blind end point (ANCOVA β [SE] = −2.836 [1.018]; p = 0.0056). Adults who only completed primary and secondary school reported greater improvements in CAARS:O-SV scores than those who completed high school and university at double-blind end point (overall p value for educational level: p = 0.0079). Higher baseline CAARS:O-SV scores were associated with greater improvements in the total CAARS: O-SV score at double-blind end point (ANCOVA β [SE] = −0.403 [0.073]; p b 0.0001). All OROS®-MPH treatment groups (18, 36 and 72 mg) experienced greater reductions (improvements) in CAARS:O-SV total scores at double-blind end point compared with the placebo group; patients receiving 72 mg demonstrated the greatest improvement in CAARS:O-SV scores versus placebo (ANCOVA β [SE] = −6.800 [1.412]; p b 0.0001). No significant relationship was observed between compliance or the following baseline predictors and change in CAARS:O-SV scores at double-blind end point: country, history of mood or anxiety disorders, history of drug or alcohol abuse or employment status. 3.3.1.2. Open-label phase. The association between male gender and improvements in CAARS:O-SV scores was also evident at open-label end point; however, this did not reach significance (ANCOVA β [SE] = 1.880 [0.977]; p = 0.055) (Table 2). Higher baseline CAARS:O-SV scores remained strongly correlated with greater clinical reductions

Table 1 Subject baseline and demographic characteristics in the double-blind and open-label periods. Parameter

Age in years (mean ± SD) Sex, n (%) Male Eductional degree, n (%) High school Primary school Secondary school University Age at diagnosis in years (mean ± SD) Employment status, n (%) Employed Unemployed Mood and anxiety disorders, n (%) Currently active History and not active Alcohol/other substance abuse disorders, n (%) Currently active History and not active CAARS:O-SV score at baseline (mean ± SD) CAARS:O-SV change score from baseline to end point (LOCF; mean ± SD) CAARS:S-S score at baseline (mean ± SD) CAARS:S-S change score from baseline to end point (LOCF; mean ± SD)

Double-blind period

Open-label period ®

Placebo

OROS -MPH

n = 96

18 mg n = 101

36 mg n = 102

72 mg n = 102

n = 370

34.5 ± 9.6

34.2 ± 10.7

33.8 ± 10.4

33.6 ± 10.3

34.3 ± 10.3

59 (61.5)

58 (57.4)

46 (45.1)

55 (53.9)

199 (53.8)

32 (33.3) 16 (16.7) 33 (34.4) 15 (15.6) 31.4 ± 12.5

22 (21.8) 16 (15.8) 43 (42.6) 20 (19.8) 30.5 ± 13.9

27 (26.5) 18 (17.6) 40 (39.2) 17 (16.7) 29.2 ± 14.3

29 (28.4) 11 (10.8) 42 (41.2) 20 (19.6) 28.9 ± 13.7

102 (27.6) 56 (15.1) 143 (38.6) 69 (18.6) 30.2 ± 13.8

61 (63.5) 35 (36.5)

62 (61.4) 39 (38.6)

62 (60.8) 40 (39.2)

63 (61.8) 39 (38.2)

231 (62.4) 139 (37.6)

10 (10.4) 25 (26.0)

10 (9.9) 27 (26.7)

11 (10.8) 36 (35.3)

17 (16.7) 32 (31.4)

45 (12.2) 111 (30.0)

0 12 (12.5) 37.2 ± 7.09a − 7.6 ± 9.9a 51.1 ± 10.3d − 5.8 ± 11.3e

1 (1.0) 12 (11.9) 35.6 ± 6.9b − 10.6 ± 10.3b 48.5 ± 12.0a − 10.4 ± 12.9d

1 (1.0) 15 (14.7) 37.3 ± 6.9c − 11.5 ± 10.0c 51.2 ± 11.0b − 11.3 ± 12.4a

1 (1.0) 15 (14.7) 36.6 ± 6.6b − 13.7 ± 11.1b 50.6 ±11.8a − 14.4 ± 15.5f

1 (0.3) 47 (12.7) 36.7 ± 6.9 − 17.8 ± 10.8 50.3 ± 11.4 − 18.8 ± 14.4

Total group

PR-OROS®-MPH = prolonged-release osmotic-release oral system formulation — methylphenidate; SD = standard deviation; CAARS:O-SV = Conners' Adults ADHD Rating Scale Observer; CAARS:S-S = Conner's Adult ADHD Self Report. a n = 95. b n = 99. c n = 101. d n = 93. e n = 91. f n = 92.

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

557

Table 2 Predictorsa of treatment outcome as assessed by change in CAARS and CAARS:S-S scores at double-blind and open-label end points. Parameter

Older age Female gender Highest educational level attained High school Secondary school University Primary school Higher baseline CAARS:O-SV score or CAARS:S-S score, respectively Treatment group OROS®-MPH 18 mg OROS®-MPH 36 mg OROS®-MPH 72 mg Placebo Country Being employed

ANCOVA regression coefficients 1) CAARS:O-SV double-blind end point

2) CAARS:O-SV open-label end point

3) CAARS:S-S double-blind end point

4) CAARS:S-S open-label end point

− 0.105; p = 0.0358b 2.836; p = 0.0056b Ese = 0.28 p = 0.0079b 4.100; p = 0.0096c Ese = 0.42 0.725; p = 0.6285c Ese = 0.07 3.778; p = 0.0315c Ese = 0.38 0.000; – − 0.403; p b 0.0001b

NS NS

NS 3.407; p = 0.0114b Ese = 0.26 p = 0.0322b 0.796; p = 0.7167c Ese = 0.06 − 3.660; p = 0.0771c Ese = 0.27 0.408; p = 0.8666c Ese = 0.03 0.000; – − 0.343; p b 0.0001b

NS NS

p b 0.0001b − 3.592; p = 0.0114d Ese = 0.37 − 4.344; p = 0.0022d Ese = 0.44 − 6.800; p b 0.0001d Ese = 0.69 0.000; – NS NS

p = 0.0477b − 2.613; p = 0.0525d Ese = 0.28 − 3.285; p = 0.0153d Ese = 0.36 − 3.310; p = 0.0167d Ese = 0.36 0.000; – NS − 1.991; p = 0.0467b Ese = 0.21

NS

− 0.825; p b 0.0001b

p b 0.0001b − 5.742; p = 0.0021d Ese = 0.45 − 5.908; p = 0.0014d Ese = 0.47 − 8.911; p = 0.0001d Ese = 0.71 0.000; – p = 0.0487b NS

NS

− 0.586; p b 0.001b NS

NS NS

CAARS:O-SV = Conners' Adults ADHD Rating Scale (physician rated); CAARS:S-S = Conners' Adult ADHD (self-reported, short version); NS = non-significant; Es = effect size. a Only those predictors that were found to be significant are reported here. b p value is for the significance of factor for the given end point. c p value for difference versus primary school. d p value for difference versus placebo. e Effect size calculated as the difference in least squares means divided by the pooled standard deviation.

(improvements) in CAARS:O-SV scores at open-label end point (ANCOVA β [SE] = − 0.825 [0.070]; p b 0.0001). Although being employed versus unemployed at baseline failed to show any prognostic value at double-blind endpoint, being employed was associated with a greater decrease in CAARS:O-SV score compared with being unemployed at open-label end point (ANCOVA β [SE] = −1.991 [0.998]; p = 0.0467). Treatment group continued to be predictive of outcome in CAARS:O-SV score at open-label end point (overall p value for treatment group: p = 0.0477), with a significant difference reported for both the 36 and 72 mg dosage groups compared with placebo (Table 2). No significant relationship was observed between compliance or the following baseline predictors and change in CAARS scores at openlabel end point: age, educational level, country, history of mood or anxiety disorders or history of drug or alcohol abuse.

CAARS:S-S scores at double-blind end point: age, history of mood or anxiety disorders, history of drug or alcohol abuse or employment status. Country of residence was a marginally significant predictor of change in CAARS:S-S score (p = 0.0487) at double-blind end point; however, the number of subjects recruited in each country was too small to allow for any meaningful conclusions with regard to a differential effect of treatment by country.

3.3.2.2. Open-label phase. Higher baseline CAARS:S-S scores remained strongly correlated with greater clinical improvements in CAARS:S-S scores at open-label end point (ANCOVA β [SE] = − 0.586 [0.067]; p b 0.0001). No significant relationship was observed between the other baseline variables or compliance during the double-blind period and change in CAARS:S-S scores at open-label end point.

3.3.2. CAARS-self-report: short version (CAARS:S-S) 3.3.2.1. Double-blind phase. Male patients experienced larger improvements in CAARS:S-S scores at double-blind end point compared with females (ANCOVA β [SE] = 3.407 [1.339]; p = 0.0114). There was also a tendency for patients who only completed primary or secondary school to report greater improvements in CAARS:S-S scores than those who completed high school or university (overall p value for educational level = 0.0322). Patients with higher baseline total CAARS:S-S scores also had greater improvements in total CAARS:S-S scores (ANCOVA β [SE] = − 0.343 [0.061]; p b 0.0001) (Table 2). All OROS®-MPH treatment groups experienced greater improvements in CAARS:S-S scores at double-blind end point compared with the placebo group (overall p value for treatment group: p b 0.0001); patients in the 72 mg dose group experienced the greatest decrease in CAARS:S-S scores (ANCOVA β [SE] = −8.911 [1.840]; p b 0.0001) (Table 2). No significant relationship was observed between compliance during the double-blind period or the following baseline variables and

3.4. Treatment response All OROS®-MPH treatment groups had a significantly higher probability of response than placebo at double-blind end point (p ≤ 0.0033) (Table 3). Compared with placebo, the odds of response were four-fold higher in the 72 mg group (odds ratio [OR]: 4.104; 95% confidence interval [CI]: 2.217, 7.598) and 2.5-fold higher in the OROS®-MPH 18 mg (OR: 2.692; 95% CI: 1.464, 4.951) and 36 mg groups (OR: 2.478; 95% CI: 1.353, 4.538). Adults who completed high school (OR: 0.507; 95% CI: 0.263, 0.978) and university (OR: 0.416; 95% CI: 0.203, 0.856) had a lower probability of treatment response compared with those who only completed primary school (p ≤ 0.0427); patients who completed secondary school had similar treatment response to those who only finished primary school (OR: 0.933; 95% CI: 0.504, 1.727; p = 0.8247) (Table 3). No significant relationship was observed between the other baseline variables or compliance during the double-blind phase and response rates.

558

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

Table 3 Predictorsa of treatment response and CAARS:O-SV score within the normal range at double-blind end point. Parameter

Female gender Highest educational level attained High school Secondary school University Primary school Lower baseline CAARS:O-SV score Randomization treatment group OROS®-MPH 18 mg OROS®-MPH 36 mg OROS®-MPH 72 mg Placebo

Logistic regression odds ratio (95% CIs) 5) Responseb

6) Normalizationc

–

0.320 (0.179, 0.572); p = 0.0001d –

d

p = 0.0096 0.507 (0.263, 0.978); p = 0.0427e 0.933 (0.504, 1.727); p = 0.8247e 0.416 (0.203, 0.856); p = 0.0171e 0.000; – – p b 0.0001d 2.692 (1.464, 4.951); p = 0.0014f 2.478 (1.353, 4.538); p = 0.0033f 4.104 (2.217, 7.598); p b 0.0001f 0.000; –

0.949 (0.912, 0.988); p = 0.0107d p = 0.0303d 2.687 (1.137,6.351); p = 0.0243f 3.011 (1.258, 7.205); p = 0.0133f 3.516 (1.502, 8.231); p = 0.0038f 0.000; –

CAARS:O-SV = Conners' Adults ADHD Rating Scale Observer; CI = confidence interval. a Only those predictors that were found to be significant are reported here. b Response was defined as a reduction from baseline of 30% or more in CAARS:O-SV score at double-blind end point. c Normalization was defined as a score at double-blind end point within the onesided 95% CI of the scores of the non-ADHD population. d p value is for the significance of factor for the given end point. e p value for difference versus primary school. f p value for difference versus placebo.

3.5. Prediction of CAARS:O-SV score within the normal range A lower baseline CAARS:O-SV score (p = 0.0107) and treatment with all OROS®-MPH dosages compared with placebo (p ≤ 0.0243) were positively associated with achievement of CAARS:O-SV within the normal range (within 95% range of non-ADHD population) at double-blind end point (Table 3). Compared with placebo, the odds of having a CAARS:O-SV score within normal range at double-blind end point were 3.5-fold higher in the OROS®-MPH 72 mg group (OR: 3.516; 95% CI: 1.502, 8.231) (Table 3). Female gender was associated with a decreased probability of achieving CAARS:O-SV within the normal range compared with male gender (OR: 0.320; 95% CI: 0.179, 0.572; p = 0.0001). No significant relationship was observed between the other baseline variables or compliance during the double-blind period and normalization of CAARS score. 4. Discussion There is increasing awareness of the prevalence of ADHD in adulthood and of the considerable negative impact that the disorder has on multiple aspects of patients' daily lives (Biederman et al. 2006a; Fried et al. 2006). To date, the majority of research on outcome predictors has been carried out in children and adolescents with ADHD, with results generally being inconclusive. The post-hoc analysis of the LAMDA study presented here was therefore designed to explore whether certain demographic and medical history baseline characteristics and treatment variables are predictive of clinical treatment outcomes in adults with ADHD. This analysis in adults with a confirmed diagnosis of ADHD presents divergent results with regards to those baseline factors that may help to predict better treatment outcomes. Treatment group was, unsurprisingly, a significant predictor of outcome for all end points assessed (except for CAARS:S-S scores at the end of the open-label period). This observation confirms results from the original report of a positive treatment effect with OROS®-MPH in an overall population of

adults with ADHD. Higher baseline CAARS:O-SV and CAARS:S-S scores also predicted greater improvements for five out of the six end points evaluated. This observation may be due to greater potential for improvement in patients with higher CAARS scores compared with those with less severe disease. Alternatively, greater baseline severity may reflect a more homogeneous and biologically-based ADHD subtype, which may respond more strongly to dopamine-enhancing therapy; this hypothesis requires further research regarding the relation between baseline severity and biological response variables. Further assessments are warranted to understand whether higher baseline CAARS scores do indeed predict superior response to OROS®MPH therapy in adults with ADHD or if, perhaps, there are subtle differences within the LAMDA study population (i.e. imbalances in certain CAARS subscale scores between subgroups) that could help further explain this observation. Crucially, the role of baseline severity should always be interpreted in relation to the nature of the outcome assessed (Newcorn et al. 2010). In the first five analyses, outcome was defined in terms of a change from baseline. Since higher baseline CAARS scores provide greater opportunity for improvement in CAARS scores, this may explain why higher baseline scores were related to better treatment outcomes. In the final analysis, outcome was defined in terms of an absolute score, i.e. normalization (the achievement of a score within a certain low range). In this instance, a lower baseline score was found to be related to a better treatment outcome. Older age was a significant predictor for the primary study endpoint (CAARS:O-SV), but not for any of the other exploratory end points assessed in the current analysis (e.g. response rate or normalization of CAARS:O-SV). One possible hypothesis around greater CAARS:O-SV decreases in older patients may relate to the fact that the dopaminergic system appears to be among the most agesensitive neurotransmitters, and that dopamine transporters have been shown to decrease with age (Volkow et al. 1996). Additionally, dopamine receptor density has been shown to decrease with age (Kaasinen et al. 2002), which may have a similar effect in adults with ADHD as they age as striatal dopamine receptor pruning does in childhood ADHD (i.e. remission of symptoms when children enter adulthood) (Andersen and Teicher 2000). Prespecified analyses of specific age groups in future trials will help further elucidate understanding of any treatment-by-age interactions in ADHD therapy. However, further studies targeted towards understanding the functional consequences of age-induced dopamine degeneration are required. Unexpectedly, gender was observed to be a significant predictor of outcomes in three of the six end points assessed, with males reporting better results than females. Although males had slightly higher baseline CAARS scores than females (approximately one point), this only explains in part why male gender may have predicted better treatment outcomes, since a change in CAARS score of almost three points was observed during the study. This finding is contrary to previous studies in adults with ADHD, which have not reported marked gender differences in the response to MPH treatment (Kooij et al. 2004; Spencer et al. 2005; Robinson et al., 2008). One laboratory study in children comparing two doses of once-daily MPH, however, showed that girls had a superior response to boys after approximately 1.5 h post-dose, but an inferior response 12 h after dosing (SonugaBarke et al. 2007). These data suggest possible gender-related pharmacological differences in stimulant therapy for ADHD, which may impact outcomes (Sonuga-Barke et al. 2007). Results from the current analysis warrant further investigation to ascertain any between-gender biological differences in outcomes and/or external factors that may have influenced results. Patients who only completed primary or secondary school were also more likely to achieve a greater treatment response than those who completed high school or university, as assessed by CAARS:O-SV and CAARS:S-S scores. There is no obvious rationale for this observation and this finding may be due to differences in baseline

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

characteristics, even though baseline characteristics have been controlled for in the model. Specifically, it may be that patients who failed to achieve higher academic qualifications had more severe disease (Latimer et al. 2003). Indeed, patients with a higher education had lower baseline CAARS scores in the current study (data not shown). Although being employed failed to show any prognostic value at the end of the double-blind phase, results from the open-label phase of the study demonstrated that patients who were employed achieved a superior treatment response in CAARS:O-SV to those who were unemployed, implying that the double-blind phase of the study may have been of insufficient duration to examine the impact of employment on treatment outcomes. As employment was only shown to be predictive in one of six end points assessed (and not the primary end point), this observation should be interpreted with caution. While a growing number of controlled studies have now examined the efficacy of MPH in an overall population of adults with ADHD, the majority have not examined potential predicting factors of clinical response to MPH (Faraone et al. 2004; Spencer et al. 2005; Biederman et al. 2006b; Biederman et al. 2006c; Fallu et al. 2006). Moreover, those studies that have assessed this failed to establish a link between a number of predictors (age, gender, intelligence, social class and psychiatric co-morbidity) and response (Kooij et al. 2004; Spencer et al. 2005). Interestingly, the analysis reported here demonstrated that a number of baseline factors, including age, gender, education and employment status, predicted outcomes for some of the end points assessed. However, no single variable was predictive of all six end points assessed and the rationale to explain how some variables predict response is highly speculative. Therefore, further research is required to confirm these findings and examine the neurobiological and genetic mechanisms involved. Moreover, those studies that have focused on predictors of treatment outcome in children with ADHD have indicated that outcome is not associated with any particular initial variable but the interaction of a range of factors (Hechtman 1999), which may explain the inclusion of certain variables, but not others, as predictive of distinct improved outcomes in this analysis. It is envisaged that the increasing use of pharmacogenetics – the study of genetic variability in medication response – will further help to identify reliable predictors of medication response in ADHD, ultimately leading to a personalized medicine approach to ADHD (Stein and McGough 2008). The findings presented here must be seen in light of methodological limitations that primarily include a relatively short treatment period (5 weeks of double-blind plus 7 weeks of open-label treatment), which did not address long-term treatment. Furthermore, the stringent inclusion and exclusion criteria included in this study may have lacked enough variability in clinical parameters to allow for a full identification of predictors of response, thereby limiting the generalizability of results to the general population. In addition, no correction for multiple comparisons was performed as part of the analysis, although it is not expected that this would have altered the general findings. On the other hand, we have observed the treatment response in a very systematic and reliable way, and in part under double-blind conditions. For some baseline variables, the number of patients in each subgroup (e.g. patients with mood and/or anxiety, or those with alcohol or drug abuse) was quite small, which could explain why they were not predictive of outcomes. It should also be noted that compliance during the double-blind phase was very high, which probably explains why compliance did not significantly predict treatment response in any of the analyses. As such, these results need to be confirmed in more rigorous randomized, controlled clinical trials in larger patient numbers. However, despite these limitations, results of these analyses provide clinicians with valuable information on a number of baseline characteristics and treatment variables that may improve outcomes in adults with ADHD treated with OROS®-MPH. In the future, the systematic assessment and identification of potentially

559

predictive variables may contribute to the development of a theory regarding the pathogenesis of the illness, while more targeted therapy may become possible when robust predictors are identified. 5. Conclusion The identification of adults with ADHD who may respond better to pharmacological treatment will enable better targeting of therapy to positively influence outcome and improve quality of life both for patients and for their families. The results of this post-hoc analysis suggest that higher baseline severity and male gender may predict better treatment outcomes in adults receiving OROS®-MPH, although further research of either variable is required to confirm the findings and examine their neurobiological underpinnings. 7. Conflicts of interest Jan K Buitelaar has been a consultant, member of advisory board and/or speaker for Janssen Cilag BV, Eli Lilly, Bristol-Myers Squibb, UBC, Shire, Medice and Servier. J.J. Sandra Kooij has received research funding from Janssen Cilag BV and Shire and has been a speaker for Janssen Cilag BV and Eli Lilly. J. Antoni Ramos-Quiroga has received lecture and consulting fees from Janssen-Cilag and Laboratorios Rubio and research funding from Janssen-Cilag. Joachim Dejonckheere performed the analyses of this trial as a consultant on behalf of a CRO (SGS Life Sciences). Miguel Casas has received lecture and consulting fees from Janssen-Cilag and Laboratorios Rubio and research funding from Janssen-Cilag. Joop C. van Oene is an employee of Janssen-Cilag EMEA, a division of Janssen Pharmaceutica. Barbara Schäuble is an employee of Janssen-Cilag EMEA, a division of Janssen Pharmaceutica.. Goetz-Erik Trott has participated in advisory boards and provided advice to Janssen-Cilag, Medice and Novartis. He has also received honoraria for talks from Astra Zeneca, Janssen-Cilag, Medice, Novartis, Pfizer and Ratiopharm. Acknowledgements The authors would like to thank Frances Gambling of Medicus International, for her editorial assistance. Editorial assistance was funded by Janssen-Cilag. References Andersen SL, Teicher MH. Sex differences in dopamine receptors and their relevance to ADHD. Neurosci Biobehav Rev 2000;24:137–41. August GJ, Stewart MA, Holmes CS. A four-year follow-up of hyperactive boys with and without conduct disorder. Br J Psychiatry 1983;143:192–8. Barkley RA, Fischer M, Smallish L, Fletcher K. The persistence of attention-deficit/ hyperactivity disorder into young adulthood as a function of reporting source and definition of disorder. J Abnorm Psychol 2002;111(2):279–89. Biederman J, Faraone SV, Spencer TJ, Mick E, Monuteaux MC, Aleardi M. Functional impairments in adults with self-reports of diagnosed ADHD: A controlled study of 1001 adults in the community. J Clin Psychiatry 2006a;67(4):524–40. Biederman J, Mick E, Surman C, et al. A randomized, placebo-controlled trial of OROS methylphenidate in adults with attention-deficit/hyperactivity disorder. Biol Psychiatry 2006b;59(9):829–35. Biederman J, Mick E, Spencer T, et al. An open-label trial of OROS methylphenidate in adults with late-onset ADHD. CNS Spectr 2006c;11(5):390–6. Buitelaar JK, Van der Gaag RJ, Swaab-Barneveld H, Kuiper M. Prediction of clinical response to methylphenidate in children with attention-deficit hyperactivity disorder. J Am Acad Child Adolesc Psychiatry 1995;34(8):1025–32. Buitelaar JK, Ramos-Quiroga JA, Casas M, et al. Safety and tolerability of flexible dosages of prolonged-release OROS methylphenidate in adults with attention-deficit/ hyperactivity disorder. Neuropsychiatr Dis Treat 2009;5:457–66. Conners C, Erhart D, Sparrow E. Conners' Adult ADHD Rating Scales (CAARS): Technical Manual. North Tonawanda, NY: Multi-Health Systems; 1999. Fallu A, Richard C, Prinzo R, Binder C. Does OROS-methylphenidate improve core symptoms and deficits in executive function? Results of an open-label trial in adults with attention deficit hyperactivity disorder. Curr Med Res Opin 2006;22(12): 2557–66. Faraone SV, Biederman J, Spencer T, Mick E, Murray K, Petty C, et al. 2006. Diagnosing adult attention deficit hyperactivity disorder: are late onset and subthreshold diagnoses valid? Am J Psychiatry 163(10): 1720–1729; quiz 1859.

560

J.K. Buitelaar et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 35 (2011) 554–560

Faraone SV, Perlis RH, Doyle AE, et al. Molecular genetics of attention-deficit/ hyperactivity disorder. Biol Psychiatry 2005;57(11):1313–23. Faraone SV, Spencer T, Aleardi M, Pagano C, Biederman J. Meta-analysis of the efficacy of methylphenidate for treating adult attention-deficit/hyperactivity disorder. J Clin Psychopharmacol 2004;24(1):24–9. Fried R, Petty CR, Surman CB, et al. Characterizing impaired driving in adults with attention-deficit/hyperactivity disorder: A controlled study. J Clin Psychiatry 2006;67(4):567–74. Ghuman JK, Riddle MA, Vitiello B, et al. Comorbidity moderates response to methylphenidate in the Preschoolers with Attention-Deficit/Hyperactivity Disorder Treatment Study (PATS). J Child Adolesc Psychopharmacol 2007;17(5):563–80. Hechtman L. Predictors of long-term outcome in children with attention-deficit/ hyperactivity disorder. Pediatr Clin N Am 1999;46(5):1039–52. Kaasinen V, Kemppainen N, Någren K, Helenius H, Kurki T, Rinne JO. Age-related loss of extrastriatal dopamine D(2)-like receptors in women. J Neurochem 2002;81(5): 1005–10. Kessler RC, Adler L, Barkley R, et al. The prevalence and correlates of adult ADHD in the United States: results from the National Comorbidity Survey Replication. Am J Psychiatry 2006;163(4):716–23. Kooij JJ, Burger H, Boonstra AM, Van der Linden PD, Kalma LE, Buitelaar JK. Efficacy and safety of methylphenidate in 45 adults with attention-deficit/hyperactivity disorder randomized placebo controlled double blind cross over trial. Psychol Med 2004;34(6):973–82. Latimer WW, August GJ, Newcomb MD, Realmuto GM, Hektner JM, Mathy RM. Child and familial pathways to academic achievement and behavioral adjustment: a prospective six-year study of children with and without ADHD. J Atten Disord 2003;7(2):101–16. Li D, Sham PC, Owen MJ, He L. Meta-analysis shows significant association between dopamine system genes and attention deficit hyperactivity disorder (ADHD). Hum Mol Genet 2006;15(14):2276–84. Medori R, Ramos-Quiroga JA, Casas M, et al. A randomized, placebo-controlled trial of three fixed dosages of prolonged-release OROS methylphenidate in adults with attention-deficit/hyperactivity disorder. Biol Psychiatry 2008;63(10):981–9. The MTA Cooperative Group. Moderators and mediators of treatment response for children with attention-deficit/hyperactivity disorder. The multimodal treatment study of children with attention-deficit/hyperactivity disorder. Arch Gen Psychiatry 1999;56(12):1088–96.

National Institute for Health and Clinical Excellence. Attention deficit hyperactivity disorder: diagnosis and management of ADHD in children, young people and adults (Clinical Guideline CG72). Available at: http://www.nice.org.uk/CG72. Accessed on: September; 2008. Newcorn JH. Managing ADHD and comorbidities in adults. J Clin Psychiatry 2009;70(2): e40. Newcorn JH, Stein MA, Cooper KM. Dose–response characteristics in adolescents with attention-deficit/hyperactivity disorder treated with OROS® methyphenidate in a 4-week, open-label dose–titration study. J Child Adolesc Psychopharmacol 2010;20 (3):187–96. Robinson RJ, Reimherr FW, Marchant BK, Faraone SV, Adler LA, West SA. Gender differences in 2 clinical trials of adults with attention-deficit/hyperactivity disorder: a retrospective data analysis. J Clin Psychiatry 2008;69(2):213–21. Rösler M, Retz W, Retz-Junginger P, et al. Prevalence of attention deficit-/hyperactivity disorder (ADHD) and comorbid disorders in young male prison inmates. Eur Arch Psychiatry Clin Neurosci 2004;254(6):365–71. Rösler M, Retz W, Yaqoobi K, Burg E, Retz-Junginger P. Attention deficit/hyperactivity disorder in female offenders: prevalence, psychiatric comorbidity and psychosocial implications. Eur Arch Psychiatry Clin Neurosci 2009;259(2):98-105. Sonuga-Barke EJ, Coghill D, Markowitz JS, Swanson JM, Vandenberghe M, Hatch SJ. Sex differences in the response of children with ADHD to once-daily formulations of methylphenidate. J Am Acad Child Adolesc Psychiatry 2007;46(6):701–10. Spencer T, Biederman J, Wilens T, et al. A large, double-blind, randomized clinical trial of methylphenidate in the treatment of adults with attention-deficit/hyperactivity disorder. Biol Psychiatry 2005;57(5):456–63. Spencer T, Wilens T, Biederman J, Faraone SV, Ablon JS, Lapey K. A double-blind, crossover comparison of methylphenidate and placebo in adults with childhood-onset attention-deficit hyperactivity disorder. Arch Gen Psychiatry 1995;52(6):434–43. Stein MA, McGough JJ. The pharmacogenomic era: promise for personalizing attention deficit hyperactivity disorder therapy. Child Adolesc Psychiatr Clin N Am 2008;17 (2):475–90 xi-xii. van der Oord S, Prins PJ, Oosterlaan J, Emmelkamp PM. Treatment of attention deficit hyperactivity disorder in children Predictors treatment outcome. Eur Child Adolesc Psychiatry 2008;17(2):73–81. Volkow ND, Ding YS, Fowler JS, et al. Dopamine transporters decrease with age. J Nucl Med 1996;37(4):554–9.