ADHD medication

Stimulants vs non-stimulants

Every major clinical guideline puts stimulants first. The 2018 Cortese network meta-analysis is why: stimulants beat placebo on symptom reduction and functional outcomes, and beat non-stimulants on most measures. That covers methylphenidate (Ritalin, Concerta) and amphetamine salts (Adderall, Vyvanse). The same-day effect — measurable within hours of a therapeutic dose — is part of why the evidence base builds the way it does.

Mechanically: stimulants raise dopamine and noradrenaline in the prefrontal cortex, which is where working memory, inhibition, and sustained attention sit — the precise functions most impaired in ADHD. Treating the bottleneck directly is why the effect size is what it is.

Non-stimulants are second-line and work differently — atomoxetine (Strattera) is a selective noradrenaline reuptake inhibitor; guanfacine (Intuniv) and clonidine are alpha-2 adrenergic agonists. Response rates are lower, onset is slower (2–6 weeks). Pliszka (2007) places them as appropriate when stimulants are contraindicated, poorly tolerated, or when there’s comorbid anxiety or a tic disorder.

Non-stimulants are not a gentler version of stimulants. Different class, different profile. The choice should be driven by comorbidities, side-effect tolerance, and history — not by a preference for something that sounds milder.

Finding the right dose

Most adults end up above the first dose their prescriber considers adequate (see the community finding below: 68% of 211). The reason is structural — there isn’t a pharmacologically “correct” dose, only the one that works for this person at this point in time, and getting to it takes a feedback loop most prescribing schedules don’t leave time for. The National Institute for Health and Care Excellence (NICE) guideline NG87 starts methylphenidate at 5mg two or three times daily and increases weekly on response and tolerability.

The feedback loop needs data. Symptom rating scales at each review — ADHD Rating Scale, or Conners’ Adult ADHD Rating Scale — give the clinician something comparative across doses. Without one, it’s hard to tell a dose that isn’t working from a dose that’s working well on a brain that ADHD makes bad at noticing it.

What to look for, beyond “I feel better”: tasks started without extended delay, fewer re-reads on the same paragraph, less time-blindness, fewer interruptions in conversation. These are more reliable markers than a general mood read, and they’re the things a partner or colleague can confirm independently.

Side effects that usually resolve with dose adjustment: appetite suppression peaks at weeks 2–4 and is usually moderate; mild heart-rate rise of 3–5 bpm is clinically insignificant in healthy adults; emotional blunting or feeling “flat” is a signal that the dose may be too high, not an inherent property of the medication.

Tolerance and tachyphylaxis

The dose that worked well for a year starts wearing off faster, or the clarity diminishes with nothing else changed. That’s tachyphylaxis — rapid tolerance development, distinct from longer-term tolerance, where the same dose produces diminishing effects. In stimulants for ADHD it can happen within a single day (afternoon weaker than morning) or unfold over weeks and months of consistent use.

Wigal et al. (2019) put the affected proportion at 20–40%, which is high enough that most prescribers will see it within a panel. The mechanism isn’t fully worked out but likely involves dopamine-receptor downregulation in response to sustained dopaminergic stimulation. The usual clinical response is a dose increase — which may work, but it isn’t the only option, and reaching for it first can mask what’s actually going on.

The alternatives a prescriber may try when tachyphylaxis is suspected:

• Medication holidays — scheduled breaks (typically 1–2 days per week, or 1–2 weeks per year) to allow receptor resensitisation. Evidence quality is modest but the practice is common and generally well-tolerated.
• Formulation switch — moving between methylphenidate and amphetamine salts, or between immediate-release and extended-release versions. Different mechanisms can restore response.
• Augmentation — adding a low-dose non-stimulant to the existing stimulant regimen. Guanfacine is the most-studied adjunct.

Before any of those, the diagnostic question. Is this actually tolerance, or has something else changed — sleep, stress, diet, hormone phase, a new life demand? A medication that “stopped working” deserves the investigation before the dose increase.

The medication cliff

The cliff isn’t a return to baseline. It’s an overshoot below it — a brief window in the late afternoon or early evening where symptoms are distinctly worse than they were before the morning dose. Extended-release stimulants typically cover 8–12 hours depending on the formulation and the person’s metabolism; the cliff lands on the way down.

Different people get different versions. Most common is irritability and emotional reactivity. Others get an executive-function crash, hyperfocus on the wrong things, or heightened impulsivity. The timing is the part that hurts — late afternoon or early evening is exactly when family life, dinner, school pickup, and any evening commitment all want executive function at the same time.

Clinical approaches to the cliff:

• Low-dose afternoon top-up — a small dose of immediate-release methylphenidate or amphetamine (typically 25–50% of the morning dose) timed to bridge the gap before the main dose wears off completely. Prescribers vary on willingness to prescribe this; the evidence supports it for people with clear functional impairment in the evening.
• Longer-acting formulation switch — Vyvanse (lisdexamfetamine) tends to produce a smoother tail-off than osmotic methylphenidate preparations in many people, with less abrupt rebound.
• Adjusting the morning dose window — taking the morning dose 30–60 minutes earlier shifts the entire coverage window and can move the cliff past the critical evening period, though may affect sleep onset.

And the cliff matters for relationships. The rebound is often disproportionate and lands on whoever is in the room. Partners and family who don’t understand the mechanism can read it as the “real” personality, the one the medication has been concealing — which is wrong, and which is also where serious relational damage starts.

Sleep interaction

The relationship is bidirectional and not always in the direction readers expect. Kidwell et al. (2015) found stimulants associated with longer sleep-onset latency and shorter total sleep duration — so the obvious direction is real and the timing of the last dose is the primary modifiable variable. But ADHD also impairs sleep independently of medication. Delayed sleep phase syndrome — natural sleep-wake cycle shifted 2–6 hours later than conventional times — is much more prevalent in ADHD than in the general population. Some readers find sleep actually improves on stimulants because the calming of mental restlessness outweighs the onset-delay effect. The right cut-off time is the one the reader and prescriber arrive at empirically, not the one in the guideline.

Practical calibration for sleep:

• Cut-off times for stimulant doses vary widely between individuals — 12 hours before intended sleep is a common starting guideline, but some people tolerate a dose 8 hours before bed without effect, while others need 14 hours.
• Melatonin (0.5–3mg) 30–60 minutes before target sleep time is widely used and has reasonable evidence for sleep-onset improvement in ADHD populations. NICE NG87 references melatonin for sleep difficulties in children with ADHD; adult evidence is more limited but clinically accepted.
• Evening blue-light reduction and consistent wake times have outsized effects in delayed sleep phase syndrome (DSPS) — these apply whether or not medication is involved.

Taking breaks

Three distinct reasons to plan a break, and the reason determines the shape. Treat them as separate decisions, not one.

Growth and weight monitoring (mainly children and adolescents). Stimulant-related appetite suppression reduces caloric intake during the day. Weekend or school-holiday breaks have historically been prescribed to allow catch-up growth, though the magnitude of the effect is debated and varies by formulation.

Receptor resensitisation. The tachyphylaxis use case above. Typically 1–2 consecutive days off per week, or a longer break during a lower-demand period like annual leave. Most people report an uncomfortable but manageable return of ADHD symptoms during the break; that’s the cost of finding out if the response will come back.

Assessing current need. Adults stable on medication for several years may benefit from a supervised trial off it to check the diagnosis and dose still fit. Life circumstances, neurological development, and comorbidities all change. NICE NG87 already recommends annual review for adults, including whether medication remains appropriate.

What breaks aren’t for: cost-saving during expensive periods, avoiding side effects by taking medication unpredictably, or self-medicating based on perceived need that day. Inconsistent dosing without a clinical rationale makes it impossible to assess what’s working — so the next dose increase happens on bad data.