Tics Don't Block ADHD Medication: What NG87 Requires Before the First Dose

Last Saturday I sat in on a session where a room full of experienced prescribers was given a nine-year-old with confirmed ADHD, real impairment, a repaired heart defect and a mild tic — and asked what they would do. They agreed unanimously on one thing and split cleanly on another. The disagreement is the interesting part, because it exposes the single most common misconception in ADHD prescribing: that coexisting tics are a reason to avoid a stimulant. They are not.

Do tics rule out a stimulant?

No — and the instinct that they do is a hangover from older practice. NICE NG87 is clear that ADHD medication should not be withheld solely because a child also has tics. The same applies to coexisting autism.

The reasoning matters more than the rule, because it changes what you do when tics do appear to worsen after starting. Tics wax and wane on their own. That is their natural history. So if a child's tics increase three weeks into a methylphenidate trial, you genuinely cannot tell on the day whether the drug did it or whether this is simply the next peak in a fluctuating course.

If tics do clearly worsen on a stimulant, you have three options rather than one: continue and review, reduce the dose, or switch. Where tics are prominent from the outset, guanfacine has the better evidence — it can help tics in its own right — and atomoxetine does not generally worsen them. Both are reasonable, and both are decisions about which agent, not about whether to treat at all.

Why "whether" has to come before "what"

The decision to medicate has three conditions, and all three have to be met.

The second condition is the one that gets rushed. Impairment, not symptom count, is the trigger. A child can carry a confirmed diagnosis and still not need medication yet: if the school puts in enough structure, the parents have been trained, and physical activity is taking the edge off the hyperactivity, that child may be coping. You can often defer — sometimes until secondary school, when academic demand rises and the impairment finally becomes prominent. The diagnosis sets the ceiling; the impairment sets the timing.

What about consent when the parents disagree?

This is the practical trap nobody warns you about, and it has a legal answer and a clinical answer that point in different directions.

Legally, consent from one person with parental responsibility is enough to start. Practically, you want both. A live parental conflict about medication is not a good environment for the child, and where parents are separated and the child moves between households, one home giving the medication and the other withholding it produces a trial you cannot interpret and a child whose treatment is chaotic.

The mirror image is the Gillick-competent 13- or 14-year-old whose parents want medication and who does not. Whatever the legal position, a young person who does not want to take a drug will not reliably take it. As the consultant leading the session put it: ADHD medications are not life-saving medications. The decision has to be one everybody can live with.

Which drug is first-line — and why isn't it the most effective one?

NG87 sets a clear order for children aged 5 and over and young people. Each step is judged on a six-week trial at an adequate dose, though an unacceptable side effect can end a trial earlier.

Here is the part that surprises people: lisdexamfetamine is widely regarded as the more effective drug, and it is still second-line. The ordering is not about efficacy. Methylphenidate goes first because there is more accumulated clinical experience with it and because it costs less. Knowing that changes how you read a child who is doing only moderately well on methylphenidate — the next step up is a real step up.

One licensing detail worth carrying: ADHD medicines are licensed from age 6, but NICE supports off-licence use in five- and six-year-olds. Below 5, a second specialist opinion is required before medication is considered at all, and initiation is genuinely rare.

Methylphenidate is not one drug

Most of the practical skill is matching a release profile to the shape of a child's day. Immediate-release peaks in 30–45 minutes and lasts three to four hours — awkward for school, but flexible while you are finding a dose, and the trough between doses gives a child with a suppressed appetite a window in which to eat.

The modified-release preparations differ in how much of the dose is released immediately:

There is a further wrinkle that explains why afternoons are harder to cover: methylphenidate produces a degree of acute tolerance across a single day, which resets overnight. A child therefore tends to need a higher effective level later in the day than earlier — which is precisely why a Concerta-type profile, weighted towards a later release, suits the child whose concentration collapses mid-afternoon.

What has to be recorded before the first prescription?

NG87 sets a physical baseline, and there is no substitute for it. Not because the numbers are interesting on the day, but because every later monitoring visit is read against them.

• Growth — height and weight, plotted on a centile chart, not recorded as bare numbers.
• Cardiovascular — baseline pulse and blood pressure, also plotted on a centile chart, plus cardiac history and examination.
• Wider health and risk — physical health, current medicines and interactions, mental state, and in older children any risk of misuse or diversion.

The centile point is not pedantry. For an adult, 120/80 is the reference and everybody knows it. For children, normal blood pressure changes with age, and nobody can hold in their head what is normal for a six-year-old versus a fourteen-year-old. A raw reading in the notes is close to uninterpretable a year later.

On misuse and diversion, one correction worth making: the risk is not confined to the patient. A stimulant in the house is available to siblings and to parents — someone wanting to focus, or wanting the kick. Diversion is the more serious concern, and it sits mainly with immediate-release preparations. Modified-release methylphenidate and lisdexamfetamine both lower it, lisdexamfetamine partly because it is a prodrug and has to be metabolised before it does anything at all.

Which cardiac findings mean stop and refer?

Stimulants modestly raise heart rate and blood pressure. NG87 lists the findings that mean cardiology advice — usually with an ECG — before the first dose. Any one of them is enough.

Two things follow, and clinicians usually only remember the first. None of these is an automatic contraindication — each one means do not start today, get advice, and record the reason. And the flip side matters just as much: a reassuring history and examination with none of these findings does not require a routine ECG. You do not ECG every child going onto a stimulant.

The case that split the room

Leo, nine. ADHD of combined presentation, diagnosed properly by CAMHS with parent and school collateral. Despite a school support plan and a parent-training programme he cannot stay seated, does not finish work, and ran into a road last month. His confidence is dropping. The family want to start.

Three things sit in the baseline pack. A small VSD repaired at age two, discharged from cardiology and well since. A paternal uncle who died suddenly at 34, cause never established. Mild motor and vocal tics for a year — blinking and throat-clearing — waxing and waning and not distressing. And height and weight both on the 25th centile, stable, in a long-standing fussy eater.

The room agreed unanimously that you do not start on the day: the cardiac history needs a cardiology opinion and an ECG first. They split on the agent. Several argued for guanfacine, on the grounds of the tics and the low growth centile. Others held out for methylphenidate once cleared.

On the agent: stimulants are clearly more effective than non-stimulants, so they deserve every chance to work. Guanfacine's strengths are aggression, mood swings and dysregulation — and it is comparatively weak for the core inattentive symptoms that are actually impairing this child. Trading away efficacy for a problem you do not yet have is a poor bargain. Methylphenidate, once cardiology clears it, with an open conversation that the tics may worsen and will be watched.

On the growth centile: a low centile is not the concern — a falling one is. Leo has sat on the 25th consistently, which is very likely constitutional. And growth suppression matters most where growth potential is greatest: a child heading for 6'3" who ends at 6'2" has lost something measurable; a child heading for 5'6" is far less likely to see a meaningful effect. None of which removes the need to monitor growth — it removes the reason to withhold the drug.

Where your responsibility actually begins

Worth being blunt, because a proportion of prescribers now work in independent services where the commercial pressure to move quickly is real. Initiation is a specialist act. NG87 1.2.9 states that primary care practitioners should not make the initial diagnosis or start medication in children or young people with suspected ADHD, and 1.3.1 makes diagnosis a specialist act.

That is not a diminished role. After titration and stabilisation, prescribing and monitoring move into shared care — which is where most non-specialist prescribers will actually meet these children, typically under an agreement to review at least every six months. Growth, pulse and blood pressure at each dose change and six-monthly thereafter; mood and mental state at every review; appetite, sleep and general health; and, with atomoxetine, always asking directly about emergent suicidal thoughts rather than waiting to be told. That is real clinical work, it carries real risk, and it is chronically under-supported.

Knowing where that boundary sits is not a limit on your practice. The RPS framework treats recognising the limits of your own competence as a competency in its own right.

Frequently asked questions