22.3 — Dose and Response

Paracelsus wrote in the sixteenth century that "all things are poison, and nothing is without poison; the dose alone makes a thing not a poison."

It is quoted so often that it has become a cliché, and it is still the single most important sentence in pharmacology.

Water kills you at a high enough dose, by diluting your blood sodium. Botulinum toxin, the most lethal substance known, smooths wrinkles and treats muscle spasm at a small enough dose (Chapter 6.2).

There are no safe substances and dangerous substances. There are only doses.

The dose–response curve

Plot dose along the bottom and effect up the side, and you get a curve with a particular shape: shallow at first, then steep, then flattening out.

And each part of that shape has a reason.

The flat beginning — below a threshold, too few receptors are occupied to produce a measurable effect.

The steep middle — this is where small changes in dose produce large changes in effect, and it is where most drugs are used.

The plateau — once essentially all the receptors are occupied, more drug does nothing. This is the ceiling effect.

And the plateau has a very practical consequence: for a drug at its ceiling, taking more adds toxicity without adding benefit.

Paracetamol has a ceiling. Two tablets and four tablets do not differ much in pain relief, and they differ substantially in liver risk. The same is true of most NSAIDs for pain, though not for their anti-inflammatory effect, which continues rising.

Opioids largely do not have a ceiling for pain relief, which is why doses can be escalated in severe pain — and why respiratory depression also has no ceiling, which is what makes overdose fatal.

Knowing which drugs have a ceiling tells you when "take more" is pointless and when it is dangerous.

Potency and efficacy are not the same

And confusing them is one of the commonest misunderstandings about medicines.

Potency is how much drug you need. Efficacy is how much effect you can get at all.

Fentanyl is roughly 100 times more potent than morphine. It is not a better painkiller — it produces the same maximum pain relief. You simply need far less of it.

Meanwhile paracetamol and morphine differ in efficacy: no dose of paracetamol relieves severe pain the way morphine does, regardless of how much you take.

Which is why "strong painkiller" is an ambiguous phrase, and why a drug advertised as more potent is not thereby better. Potency matters for the size of the tablet. Efficacy matters for whether it will work.

And high potency has a genuine danger attached: the smaller the effective dose, the smaller the margin for a measuring error. Fentanyl's potency is exactly why contamination of street drugs with it kills so reliably — a difference of a few grains matters (Chapter 23.6).

Where a recommended dose comes from

And the answer is more empirical than people expect.

In development, a drug is tested across a range of doses in phase 2 trials (Chapter 22.15). You are looking for the dose that sits in the sweet spot: high on the effect curve, low on the harm curve.

Because there are two curves, not one. The therapeutic curve and the toxicity curve, and the gap between them is the therapeutic window (Chapter 22.1).

The recommended dose is chosen to sit inside that window for most people.

Which contains an admission worth noticing: "most people". The standard dose is a population average, and individuals vary in absorption, metabolism, body size, kidney function and receptor sensitivity. Which is why doses are adjusted, and why some drugs are monitored by blood level rather than by assumption.

Doses by weight — used in children, and for drugs with a narrow window. Children are not small adults, and dosing by weight or body surface area is standard in paediatrics.

And a specific safety point: paediatric doses are calculated per kilogram, and a decimal point error is the classic route to a tenfold overdose. Which is why paediatric prescriptions are double-checked as routine.

Maximum doses and why they exist

Not because the drug stops working, but for one of three reasons.

Toxicity. Paracetamol's 4 gram daily maximum in adults exists because of liver damage (Chapter 22.5), and the margin is narrower than people assume — particularly with low body weight, malnutrition, or regular alcohol.

The ceiling effect. No further benefit, so no reason to accept further risk.

Cumulative dose limits. Some drugs have a lifetime total, not a daily one.

Doxorubicin, a chemotherapy drug, damages the heart cumulatively, so the total received over a lifetime is tracked. Amiodarone accumulates in tissues. Bisphosphonate treatment is periodically paused for similar reasons (Chapter 21.6).

Loading and maintenance

A loading dose fills the tank. A maintenance dose keeps it there.

Because reaching a steady level takes 4 to 5 half-lives (Chapter 22.2), and sometimes you cannot wait.

Serious infection is the clearest case: a large first dose of an antibiotic reaches an effective level immediately rather than in a day.

Also loaded: digoxin, amiodarone, phenytoin, and anticoagulants in some situations.

And this is why an apparently large first dose on a prescription is frequently correct rather than an error — though it is also exactly the situation where checking is worthwhile.

Titration

Starting low and increasing gradually.

Done for three distinct reasons, and knowing which one applies tells you how fast you can go:

To let side effects settle. SSRIs and gabapentin are titrated because the initial effects fade with continued use (Chapter 20.7).

To find the individual's dose. Blood pressure drugs, thyroid replacement, insulin — the target is a measurement, not a dose.

To avoid a dangerous reaction. Lamotrigine is escalated very slowly specifically because rapid increases raise the risk of a severe rash (Chapter 20.2). This is the category where the schedule is not negotiable.

And tapering on the way down, for beta blockers, steroids, opioids, benzodiazepines and antidepressants — because of rebound, dependence or adrenal suppression.

Timing, and why some drugs have a preferred time of day

Statins were traditionally taken at night, because cholesterol synthesis peaks then and the older statins were short-acting. The longer-acting ones can be taken any time, which matters because the best time to take a tablet is the time you will actually remember.

Diuretics in the morning, for the obvious practical reason.

Steroids in the morning, to match the body's own cortisol rhythm and reduce suppression of the adrenal glands (Chapter 12.4).

Levothyroxine on an empty stomach, which usually means first thing.

Some blood pressure drugs at night — and the evidence on whether night dosing improves outcomes has gone back and forth, with the honest position being that taking them consistently matters far more than the hour.

Sedating drugs at night; stimulating ones in the morning.

Adherence — the biggest variable of all

And it deserves a section, because it outweighs most of the pharmacology above.

Around half of people with long-term conditions do not take their medication as prescribed.

Which means the commonest reason a treatment fails is that it was not taken — a fact that is frequently missed before doses are escalated or drugs are switched.

Why people stop, in roughly the order that matters:

Side effects, frequently not reported because they feel embarrassing or trivial to mention.

No felt benefit. Blood pressure and cholesterol treatment produce nothing you can feel, and prevent something invisible. This is a genuinely hard thing to sustain, and it is not irrational.

Complexity. The more tablets and the more times a day, the worse adherence gets, reliably.

Cost.

Beliefs about medication, including the concern about becoming dependent on tablets, which is worth addressing rather than dismissing.

Simply forgetting.

What helps: once-daily dosing where possible; combination tablets; linking doses to an existing routine; pill organisers; and asking about side effects directly rather than waiting to be told.

And a useful reframing for anyone taking long-term preventive medication: the drug is not treating how you feel today. It is buying a lower probability of a stroke, a heart attack or a fracture at some point in the future. That is a real thing to be buying, and it is easier to sustain when it is stated plainly.

The two things worth carrying

More is not better once you are on the plateau.

And the dose is where safety lives. Almost every medicine in this Part is safe at one dose and harmful at another, which means the instructions on the label are not bureaucracy — they are the boundary of the window.

What the next page fixes

Chapter 22.4 covers how drugs are given — tablets, injections, inhalers, patches and the rest, and why the route changes everything about how a drug behaves.