Statistics from Altmetric.com
As a unit, the kiloton is most often linked to the explosive yield of nuclear weapons in tons of TNT (trinitrotoluene, an explosive). For drugs, we are more comfortable with milligrams, a unit one million million times smaller. With paracetamol, at a population level, the kiloton unit may be more appropriate: in Europe, paracetamol sales range from under 200 tons in Greece and Portugal to 6300 tons in the UK and 10 000 tons in France.1 On a per capita basis the range is 4–5 tons per million to 30–50 tons per million. These figures relate to the total population exposure, including over-the-counter (OTC) sales, though prescribing of paracetamol is not negligible. In the UK in 2014, 42 million paracetamol-containing medicines were prescribed at a cost of £191 million,2 and 200 million packs sold OTC.3
Paracetamol was discovered in the 1950s,4 and the general view that it is effective and safe has led to such widespread use. Yet there has been considerable uncertainty over how it works. Only recently has it been generally accepted that it inhibits COX-1 and COX-2 isoenzymes, and is in fact a weak nonsteroidal anti-inflammatory drug (NSAID).5
The last few years have seen paracetamol and other commonly used drugs subjected to greater scrutiny, particularly the sharp focus of evidence-based medicine; 2015 and 2016 have seen pivotal new evidence appear. The picture is not quite as rosy as the general view would have us think.
In terms of efficacy, an overview review that examined paracetamol efficacy across acute and chronic conditions6 has been complemented with more recent data. What we now know about paracetamol and pain is that:
Paracetamol at doses between 500 and 1000 mg is in the least effective quartile of drugs for treating acute postoperative pain.7
Paracetamol 1000 mg has modest efficacy in migraine and tension-type headache.
Paracetamol at doses up to 4000 mg daily is practically ineffective in arthritis. Though marginally better than placebo, paracetamol has little chance of achieving clinically meaningful benefit in osteoarthritis.8 ,10–12
No review evidence that paracetamol works for dysmenorrhoea, neck pain, rheumatoid arthritis or cancer pain.
We have considerable evidence that as well as not being particularly effective, neither is it particularly safe.
A systematic review of observational studies found that compared with people not taking paracetamol, paracetamol use, especially at higher doses, was associated with increased mortality, cardiovascular adverse events (fatal or non-fatal myocardial infarction, stroke or fatal coronary heart disease), gastrointestinal adverse events (gastroduodenal ulcers and complications such as upper gastrointestinal haemorrhage) and estimated glomerular filtration rate decrease of at least 30 mL/min/1.73 sq m.13
Acute liver failure leading to registration for transplantation was twice as common in non-overdose paracetamol-exposed patients than with NSAIDs in a large case-population study.14
In clinical trials in chronic pain, patients taking paracetamol were four times more likely to have abnormal results on liver function tests than those taking placebo.8
Reports of patients with any adverse event in acute pain studies were the same for paracetamol (up to 1000 mg) and placebo.17
Oral paracetamol is rapidly absorbed, with maximum plasma concentrations achieved by 30 min in the fasted, but not fed, state.18 Rapid absorption of NSAIDs is now recognised to be important for promoting good analgesia, though the precise mechanism is uncertain. That may be why paracetamol has some analgesic action in acute pain.7 But only a minority of people with acute pain or headache will have good pain relief with paracetamol, especially compared with fast-acting formulations of NSAIDs. For chronic pain we have evidence of absence of any clinically useful effect of paracetamol, either alone or in combination, and at doses of up to 4000 mg daily. There may of course be circumstances where paracetamol might be useful, in paediatric pain, for treating patent ductus arteriosus, or for intravenous use during surgery.
Balanced against this at-best-modest analgesic efficacy we have considerable evidence of harm, including all the concerns that also affect NSAIDs, but with liver failure and (possibly) renal failure added. These are dose-related; more risk of harm at higher doses.
If paracetamol were an unimportant and little-used drug, none of this might amount to much. But it is not. It is a drug whose use is measured in thousands of tons, has little or no effect in many conditions and has significant adverse events. Maybe it is time to consider the evidence again and think about the possibility of change.
Patients with pain regard only large reductions in that pain to be relevant, and large reductions in pain are accompanied by major improvements in associated symptoms, including quality of life, function and work, and lower costs to health services.19 ,20 We know that only a minority of patients have adequate pain relief on any particular drug.21 We know that the majority of people with chronic pain still have moderate or severe pain while on treatment, indicating that treatment to be inadequate.22 We know that switching to another medicine in that situation results in major benefits for many.23
The future for treating pain lies in new thinking about how we deal with analgesic failure, rather than remorselessly continuing to use kilotons of a drug that we know will fail and may do harm.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.