Search IconSearch
April 12, 2016/Neurosciences/Research

Dose Response to Psychostimulants in ADHD: The Case for Vigilant Monitoring

Study shows how idiosyncratic dose-response patterns can be


By Michael Manos, PhD; Eric Geyer, BA; Ralph D’Alessio, BA; Kimberly Giuliano, MD; and Michael Macknin, MD


Cleveland Clinic is a non-profit academic medical center. Advertising on our site helps support our mission. We do not endorse non-Cleveland Clinic products or services. Policy

Substantial evidence from clinical trials confirms that psychostimulants improve symptoms in children with attention deficit hyperactivity disorder (ADHD) and that their pharmacokinetic properties are well-documented.1

There remains, however, a dearth of information on how behavior changes during stimulant titration.

What influences dose response?

The prescribe-and-wait method, in which the physician uses the skill and art of medical practice to determine the best dose, is common and may be informed only by the physician’s personal experience with patients.2 In clinical practice, however, stimulant dose response varies considerably across ADHD subtypes, patient age and gender, type of stimulant, and other variables that have been only peripherally studied.3

To date, factors that inform and predict response to stimulants remain elusive. Because the relationship between dose and response tends to be predominantly linear (as dose increases, behavioral improvement increases), clinicians have a tendency to expect this pattern and subsequently may not systematically observe and measure response as it occurs. Response variability often goes unnoticed,4 and when it is overlooked, optimal treatment can be compromised.

Early clinical practice parameters suggested that dosing should be carried out according to gross body weight, beginning with 0.3 mg/kg in a twice-daily regimen and titrated upward until undesirable side effects emerged or behavior improved. This method was untenable, however, because pharmacological factors such as drug absorption, metabolism and excretion rates produce great interindividual variability.5

Six patterns of individual dose response

Although a linear dose response is found consistently at the group level of analysis, individual children vary considerably in behavior change across dose levels.6 In addition to a linear response, there are five other patterns of response in individuals:

  • A curvilinear response is incremental improvement at low and moderate doses followed by a decrement at a high dose.
  • A threshold response is the absence of observed improvement at a low or moderate dose with abrupt therapeutic benefit at a high dose.
  • A variable response is marked by irregular effects in the wake of dose changes, such as improvement at a low dose, deterioration at a moderate dose and improvement again at a high dose.
  • A placebo response is when an individual shows equal reaction to medicine and a nonmedicine placebo capsule.
  • No change may be seen during standard dosing in some individuals who just do not respond to pharmacotherapy.


Three curves are predictable in that they tend to yield results in specific regimens and have a readily recognizable course of action. Behavior that improves as dose increases (linear response) indicates to the physician that an optimal dose may be achieved by ramping up the dose. If behavior deteriorates at higher doses (curvilinear response), the best dose is achieved by using the dose just prior to deterioration. If no response is observed at the initial doses, a further increase in dose may still yield optimal response (threshold response).

The remaining three curves, however, are unpredictable and do not help guide the physician’s next action during titration. Curves that inconsistently improve response despite increase or decrease in dose (variable response) give little guidance as to the next action, such as to terminate, change or extend the titration. The same case applies when individuals respond to treatment regardless of the medicine or dose (placebo response). Finally, some children do not respond to medicine except at unusual dose levels (no change), and these cases are seldom pursued in primary care. Dose curves are illustrated in the figure.

15-NEU-2530 Inset

Figure. Six dose-response curves from a sample of youths ages 5 to 17 prescribed methylphenidate, mixed amphetamine salts (MAS) or sculpted dosing of MAS (see main text) in Cleveland Clinic’s ADHD Medication Monitoring Clinic. A lower score indicates symptom reduction. ASQP = Abbreviated Symptom Questionnaire for parents. The dose response of the 249 youths was as follows: linear, n = 106; curvilinear, n = 65; threshold, n = 18; variable, n = 19; placebo, n = 21; no change, n = 20.

Real-world investigation of dose-response patterns

These distinctions describe dose-response patterns that confound titration of psychostimulants during pharmacotherapy for children with ADHD. They also identify characteristic patterns with implications for pediatric practice for ADHD treatment. We investigated these patterns in the Medication Monitoring Clinic of Cleveland Clinic’s ADHD Center for Evaluation and Treatment.

A sample of 249 youths meeting DSM-III or DSM-IV diagnostic criteria for ADHD was evaluated using a four-week, double-blind, placebo-controlled protocol. Physicians prescribed twice-daily dosing of methylphenidate, once-daily dosing of mixed amphetamine salts (MAS) or sculpted dosing of MAS (higher dose in the morning, 5-mg dose at about 3 p.m.) for youths ages 5 to 17 years. Data were collected at baseline and across the pharmacological protocol, which specified four possible dosing sequences (BL = baseline, A = placebo, B1 = 5 mg, B2 = 10 mg, B3 = 15 mg):

  • BL, A, B1, B2, B3
  • BL, B1, A, B2, B3
  • BL, B1, B2, A, B3
  • BL, B1, B2, B3, A

Each participant was randomized to one of the four sequences. Teachers and parents rated symptom presence at baseline and at the end of each treatment week using the Conners Abbreviated Symptom Questionnaire for parents.7


Children’s responses to individual psychostimulant medications varied by age. Predictable dose-response curves (linear, curvilinear and threshold) were associated with younger children (< 10 years), and unpredictable curves (variable, no response and placebo) were associated with older children (> 9 years). The idiosyncratic nature of children’s responses to psychostimulants requires close monitoring during titration to optimize treatment. These results underscore the need to use more sophisticated dosing strategies at older ages.


  1. Subcommittee on Attention-Deficit/Hyperactivity Disorder; Steering Committee on Quality Improvement and Management. ADHD: clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity disorder in children and adolescents. Pediatrics. 2011;128(5):1007-1022.
  2. Manos M. Nuances in initiating stimulant treatment for children with ADHD. In: Rostain A, ed. The Year In ADHD. Philadelphia: Current Medicine Group; 2009:1-4.
  3. Manos MJ. Pharmacologic treatment of ADHD: road conditions in driving patients to successful outcomes. Medscape J Med. 2008;10(1):5.
  4. Rapport MD, Stoner G, DuPaul GJ, et al. Methylphenidate in hyperactive children: differential effects of dose on academic, learning, and social behavior. J Abnorm Child Psychol. 1985;13(2):227-243.
  5. Rapport MD, DuPaul GJ, Kelly KL. Attention deficit hyperactivity disorder and methylphenidate: the relationship between gross body weight and drug response in children. Psychopharmacol Bull. 1989;25:285-290.
  6. Dupaul G, Stoner G. ADHD in the Schools. New York: Guilford Press; 1994.
  7. Margalit M. Diagnostic application of the Conners Abbreviated Symptom Questionnaire. J Clin Child Psychol. 1983;12(3):355-357.

Dr. Manos is Director of Cleveland Clinic’s Center for Pediatric Behavioral Health.

Mr. Geyer is a psychometrist and Mr. D’Alessio a clinical research assistant, both in the Center for Pediatric Behavioral Health.

Drs. Giuliano and Macknin are staff physicians in the Department of General Pediatrics.


Related Articles

illustration of an alzheimer brain and a packet of sildenafil pills
March 11, 2024/Neurosciences/Research
Sildenafil as an Alzheimer’s Candidate Drug: Further Support From Insurance Database and Mechanistic Studies

Real-world claims data and tissue culture studies set the stage for randomized clinical testing

June 15, 2022/Neurosciences/Research
MINDS Study Will Assess Lifestyle Interventions for Slowing Brain Pathology in Preclinical Stages

New grant-funded investigation illustrates impact and reach of Cleveland Clinic Brain Study

March 9, 2022/Neurosciences/Research
A Closer Look at the Cleveland Clinic Brain Study

How the new longitudinal investigation could become the Framingham Heart Study of brain health

20-NEU-1959738 powered-exoskeleton-for-gait-training_650x450
October 23, 2020/Neurosciences/Research
Study Signals Potential of Powered Exoskeleton for Gait Training in Multiple Sclerosis

Pilot findings show good patient acceptance and safety, early hints of efficacy

October 21, 2020/Neurosciences/Research
Electronic Screening for Sleep Disorders Among Healthcare Workers Is Feasible – and Sorely Needed

Study finds high prevalence of symptoms, willingness to seek treatment

20-NEU-1946228 neurofilaments_650x450
September 9, 2020/Neurosciences/Research
Serum Neurofilament Light as a Progressive MS Biomarker: Guideposts for the Road Ahead

Panel outlines research priorities around a promising alternative to imaging markers

20-NEU-1951118_Impella-temporary-LVAD_650x450_ (1)
September 4, 2020/Neurosciences/Research
Study Finds Clues to Acute Neurologic Events With Short-Term Impella Cardiac Support

Suspected factors include antithrombotic intensity, time on device, presence of thrombocytopenia

20-NEU-1929179 deep-brain-stimulation-for-TBI_650x450
September 3, 2020/Neurosciences/Research
Could DBS of the Cerebellar Dentate Nucleus Enhance Post-TBI Rehabilitation?

Preclinical studies will assess whether method developed for stroke recovery curbs deficits after brain injury