The Strongest Opioid Painkillers: What Potency Actually Means

In Brief

A list like this reads as a ranking: strongest at the top, weakest at the bottom, the rest in between. The numbers are real, taken from a network meta-analysis of 52 randomised trials covering 16 opioids (PMID 36523208).

To use them, though, you have to know what was measured. And what was measured is unexpected.

What Was Actually Measured

All the data come from one device — a pump the patient operates. After surgery a person is given a button: the pain rises, they press, the machine delivers a dose intravenously and then locks for a set time.

The authors did something that makes the work worth reading: they refused to use existing conversion tables. Opioid potency is usually taken from reference tables that accumulated over decades and were copied from source to source. Instead, potency here was computed as the inverted ratio of total consumption — how much of the drug a person gave themselves before the pain settled to a comparable level.

Which gives a simple rule for reading the table: 423 for sufentanil means about 423 times less of it was needed by weight, not that it relieves pain 423 times better. The relief achieved across the compared trials was comparable; that was a condition of inclusion.

So what was measured is not a property of the molecule in isolation, but the behaviour of a person with a button. That matters for what follows.

Three Caveats the Authors Make Themselves

• This is intravenous only. Every trial used a pump in acute pain. The ratios do not transfer directly to a patch, a tablet or a sublingual form.
• Short-acting drugs look weaker. The authors write it plainly: short-acting opioids are less potent compared with longer-acting drugs, morphine for example, probably because of the shorter intervals for readministration. Part of the order in this table is about the clock, not about power.
• The results in places contradict the literature. That is not an admission of weakness: reference tables have long been criticised for imprecision, and here they were deliberately not used. The disagreement follows from the method rather than from an error.

Why This Is Not a Ranking of «Which Is Better»

In August 2026 the French national guideline on the appropriate use of opioids was published (PMID 42371399). It contains a sentence that devalues any ranking built on potency:

«Regardless of their pharmacological potency, the benefit–risk balance of opioid analgesics is primarily dose-dependent, with risks, including the development of opioid use disorder and overdose, shared across all opioids.»

And immediately afterwards the authors say something still more awkward for lists: the classification of opioid analgesics proposed by Lussier and Beaulieu is preferable to the WHO three-step ladder with its division into weak and strong opioids.

So a current guideline disputes not the particular numbers but the habit of arranging opioids by strength. A strong drug is not a better or a worse drug. It is a drug you need less of by weight.

What Potency Means in Practice

A difference in potency settles a question of dosing, not a question of choice.

When a drug is tens of times stronger than morphine, the work is done in micrograms rather than milligrams, and the cost of an arithmetic slip grows by the same factor. That is why the strongest opioids live where there is supervision, not in a bathroom cabinet.

The choice between drugs is decided by other things: how fast relief is needed, how long it must hold, the state of the kidneys and liver, what else the person is taking, whether they have had opioids before. The potency table answers none of those questions.

Where These Drugs Actually Exist

Checking the Italian medicines register shows that the table and reality part company more sharply than one might expect.

The strongest drug on the list does not exist outside a hospital. Every registered sufentanil pack is classed for hospital supply. A sublingual form for use outside one did exist in Europe: Dzuveo was authorised across the EU on 25 June 2018 — and even that is prescription-only, given by a healthcare professional experienced in treating pain, in a monitored setting, and never for more than 48 hours. A second such product, Zalviso, had its authorisation withdrawn by the European Commission on 5 August 2022 at the request of the marketing authorisation holder, for commercial reasons.

Fentanyl looks different: almost a hundred registered names, and mostly patches — a form for pain that lasts, not for a shot on demand.

Buprenorphine has split by purpose, and the supply category shows it: patches are dispensed as an analgesic, sublingual forms as treatment for opioid dependence. One drug, two different uses, and the potency table cannot see the difference at all.

Risk Follows the Dose, Not the Line in the List

Here the European picture has a feature worth naming, because it is often replaced by the American one.

France is among the least affected OECD countries for opioid mortality: an estimated 300–400 fatal overdoses a year, linked primarily to methadone (40%) and heroin (20%), with analgesic opioids far behind at 12%. For comparison, in the United States in 2017 opioids were implicated in two-thirds of 71,568 overdose deaths.

But to stop at that figure is to stop reading too early. The guideline's concern lies elsewhere: a new profile of the person who overdoses has appeared. Not the «typical» drug user, but someone opioid-naïve or an occasional user — someone whose dependence began with prescribed treatment.

That is why the sentence about dose matters more than the table about potency. The danger is not created by which opioid was prescribed, but by how much, for how long, and under what supervision.

And What About Tablets

Everything above concerns intravenous administration. For tablets the counting is done differently, and the figures come out different.

The usual instrument is a morphine-equivalents chart. On it, 7.5 mg of oral hydromorphone is treated as equal to 20 mg of oral oxycodone — because both are counted as equal to 30 mg of oral morphine. It is worth noticing how that equality was arrived at: the two drugs were never compared with each other, both were converted through morphine. And it comes not from a trial but from a teaching chart for emergency physicians (PMID 37465775), which says in as many words that the value is extrapolated from the chart rather than measured head to head.

And here the current guideline says something unexpected. Verbatim: there are no universally accepted opioid conversion tables for acute pain. The recommendation is to work from each drug's own product information, and when rotating from one opioid to another to reduce the calculated dose by 25–50%, because people respond to opioids differently and tolerance does not carry over completely between them.

So the chart that yields those tidy equalities is, in the guideline authors' view, not fit to serve as an exact measure — it is a starting point, not an answer.

And one last thing, which closes the question of rankings altogether. From the same guideline: no clear superiority of «strong» over «weak» opioids has been demonstrated, and systematic reviews have found no significant differences in efficacy between opioid drugs — neither in pain relief nor in the recovery of daily function.

That makes a third way of counting and a third set of numbers. Intravenously one set of ratios, by the equivalents chart another — and the guideline, standing over both, says the difference does not translate into outcomes for the person taking them.

Tapentadol: The Case Where «How Strong» Has No Answer

The third drug usually placed beside the first two shows the disagreement most clearly. The guideline describes the evidence on it, in its own word, as heterogeneous: some meta-analyses report superiority over oxycodone, others non-inferiority, and one comparative analysis found it less effective than tramadol. Attaching a potency figure to it would invent a precision the sources do not contain.

What it does have is something that is established. Systematic reviews found no differences in the frequency of adverse events between opioids with one exception — tapentadol, which shows a lower incidence than oxycodone. Four pieces of work support that, the Cochrane review among them.

And the counterweight straight away, without which the picture would be flattering: in trials running twelve weeks, treatment was discontinued for lack of efficacy in 6.2% of patients on tapentadol against 3.4% on oxycodone. The authors call both proportions low, and that is a necessary part of the sentence: 6.2 against 3.4 cannot be read as «twice as bad», the difference rests on small numbers. But the direction is reversed: better tolerated, and more often abandoned for not working well enough.

Which is the answer about this drug: «how strong» is a question the data do not settle, while «how well tolerated» is one they do.

The Bottom Line

The numbers in the table are real and were obtained by a good method — possibly the best available, because the authors declined to copy the old reference works and counted again from the trials.

But they answer a narrow question: how much drug a person with a button needed after surgery. Not «which relieves pain better», not «which is safer», and certainly not «which to choose».

One practical rule follows, and it runs against intuition: the higher a drug sits on the list, the less of it is needed and the less room there is for error. Potency is not might, it is scale. And where the scale is finer, supervision matters more.

This article is for information. Opioid analgesics are used only as prescribed by a physician; it contains no guidance on doses, regimens or conversion from one drug to another, and none should be inferred from it.