The medication cliff

What tachyphylaxis is, and is not

Tachyphylaxis in pharmacology is rapid loss of drug response after repeated administration. In adult ADHD prescribing, the term names a sustained loss of stable-state response to stimulant medication over weeks to months, in the absence of under-dosing, adherence problems, or comorbidity. Handelman & Sumiya 2022¹ is the cleanest current literature synthesis; Ross, Fischhoff & Davenport 2002² first formalised the term in adult ADHD with a case series. Some clinicians deny the phenomenon exists at therapeutic doses; others treat it as universal. Both positions are wrong. Real for some, artefact for others.

Four things tachyphylaxis is not. It is not the within-day wear-off of a short-acting stimulant — that is the expected duration of action. It is not the 2–4 hour post-dose rebound — that is a separate phenomenon with a different treatment. It is not the fading of the first-month “this is incredible” feeling while the objective response is stable — that is placebo-response decay. And it is not seasonal or life-event-related fluctuation in response — that is the differential, below.

Placebo-response decay — the most common “tachyphylaxis”

Stimulant RCT placebo arms reliably show a 20–30% symptom improvement that fades over 6–12 weeks of follow-up. That fade is the single most common explanation for the “stopped working” complaint a prescriber sees.

The pharmacology underneath. The patient who started a stimulant a month ago and felt dramatically better in week one was experiencing the drug effect plus an expectancy and novelty component layered on top. As the expectancy fades over weeks 4–12, the subjective experience drops even though the drug’s pharmacological effect is stable. The patient interprets the drop as the drug failing. An objective rating-scale measurement — ASRS, AISRS, or CGI-S administered at post-titration baseline and again at the complaint visit — typically shows preserved response, confirming artefact.

Operationally: the first move at the “stopped working” complaint is the rating-scale comparison, not a dose increase. If scores are preserved, the explanation is artefact, the right move is reassurance and continued treatment at the current dose, and the patient is at risk of escalating treatment changes against the wrong target.

Mechanism candidates — three, all mechanistic

Three candidates carry the strongest mechanistic basis for real tachyphylaxis; none has been proven causal in adult ADHD.

D2/D3 receptor downregulation. Chronic exposure to elevated synaptic dopamine produces postsynaptic receptor downregulation in preclinical models. Volkow et al. 2009³ PET imaging in stimulant-exposed populations shows altered D2/D3 availability in striatal regions. The inference — chronic stimulant treatment shrinks the post-synaptic response over time — is plausible, but the human evidence specific to ADHD tachyphylaxis is limited.

Presynaptic D2 autoreceptor sensitisation. Increased autoreceptor sensitivity would reduce dopamine release per stimulant exposure. Preclinical support; clinical translation thin.

Vesicular monoamine transporter regulation. Chronic stimulant exposure may alter VMAT2 expression, changing vesicular dopamine storage and release dynamics (Fleckenstein et al. 2007⁴). Mechanistically plausible; clinically unverified in this context.

The eight differentials

Before calling it tachyphylaxis, an adult prescriber should rule out eight other things. The reader needs the list to have the right conversation.

• Sleep deprivation. ADHD-related sleep dysregulation is common; six-hour nights produce a worse stimulant response than eight-hour nights at the same dose (Hvolby 2015⁷). The prescriber’s first question should be the sleep question.
• Comorbid depression. Stimulants do not treat depression. Pre-existing depression that surfaces months into stimulant treatment is common (Biederman et al. 2008⁸); the depressed reader will experience the depression as “the drug stopped working.” Patient Health Questionnaire (PHQ-9) at the visit.
• Menstrual cycle interaction. Luteal-phase reduced stimulant response is documented (Quinn & Madhoo 2014⁹). Cyclical pattern is the diagnostic clue. Full treatment at cycle interactions.
• Perimenopause. Estrogen drop, dopamine drop, the dose that worked at 38 underperforms at 46. Full treatment at perimenopause and ADHD.
• Life stressor or cognitive load increase. New job, divorce, caregiving role, parenting transition. The drug effect is constant; the demand exceeded the capacity even with treatment.
• Dose drift via weight or habituation.30 mg at 55 kg is a higher mg/kg dose than 30 mg at 70 kg. Weight gain over months can move a patient out of the effective range without a perceived “change” in dosing.
• Sleep apnea, hypothyroidism, iron or vitamin D deficiency.Each independently affects cognitive function. Worth screening at the “stopped working” complaint in otherwise medically unmonitored patients.
• Adherence drift. Missed doses, irregular timing, food-effect changes, generic-to-generic batch differences. The patient often does not register adherence drift as a change; the prescriber should ask the specific timing questions.

Tachyphylaxis vs rebound — keep them separate

Rebound is the 2–4 hour post-last-dose window where symptoms transiently worsen — irritability, emotional dysregulation, fatigue — as plasma catecholamines drop rapidly. Documented in pediatric work and observed in adults (Wigal 2009¹⁰); time-locked to dose pharmacokinetics. Tachyphylaxis is a sustained decline in stable-state response. The same patient can have both; the clinical move differs. Rebound is treated with formulation extension, a short-acting booster to soften the tail, or an α2A agonist at bedtime. Tachyphylaxis is treated with the differential workup above and the graded response below.

The graded clinical response

No major guideline gives an algorithmic protocol. Clinical convention across NICE NG87¹¹, APSARD 2024¹², and the Handelman & Sumiya 2022 review converges on a graded sequence.

• Confirm by measurement. Rating-scale comparison against the post-titration baseline. If scores are preserved, the explanation is artefact and the sequence stops here.
• Rule out the eight differentials. Sleep, depression, cycle, perimenopause, life stressor, dose drift, medical contributors, adherence.
• Dose titration upward. If headroom exists — the patient at 30 mg Vyvanse can go to 40, 50, 60, 70 — this is the first medication change.
• Formulation switch within class. Adderall XR to Vyvanse; Concerta to Focalin XR. Different release profile, sometimes resolves without changing the molecule.
• Class switch. Methylphenidate to amphetamine or vice versa. Roughly 30% of patients who fail one class respond to the other (Faraone 2018⁶).
• Drug holiday. Short for assessment, longer for receptor resensitisation. Evidence base for adult holidays is thin; clinical convention only.

The reflex move — “switch me to something else” without the differential workup — is the failure mode APSARD 2024 identifies. The patient who has already had three switches and still feels the same is usually the patient who needed one of the eight differentials addressed.

The honeymoon reframe

“Honeymoon period” is the most common patient framing of the first-month dramatic response followed by perceived decline. The framing earns its space because the reader uses it. The pharmacological picture is often the reverse: the dramatic first-month response reflects the contrast against decades of untreated ADHD, the expectancy component, and the genuine therapeutic effect on a never-treated system. The post-honeymoon state — drug working at the expected Cortese 2018⁵ amphetamine effect, expectancy effect faded, baseline reset to the new normal — is often the actual stable therapeutic state.

The right diagnostic question is not “how do I get the honeymoon back” but “is what I am experiencing now the drug working as expected, or is something specific declining.” The eight differentials answer the second version.