Fentanyl Carries Triple the Respiratory Risk of Codeine, Hospital Study Finds

Patients given fentanyl in the hospital were more than three times as likely to develop dangerously slowed breathing as patients given codeine, according to a study of 32,909 adults published in BMC Medicine. The analysis, led by researchers at the University of Manchester, is among the first to rank individual opioids by respiratory risk using records of what was actually administered at the bedside rather than what was prescribed on paper.

Respiratory depression is the mechanism by which opioids kill. Breathing slows, oxygen levels fall, and without intervention the patient stops breathing. Clinicians have long known the risk varies by drug and dose, but the comparative evidence has been thin enough that guidelines mostly treat potent opioids as interchangeable once converted to a common dose scale.

What the researchers measured

The team drew on electronic health records from Salford Royal Hospital in Northwest England, covering adults admitted between September 2014 and October 2017 who received opioids for non-cancer pain. Rather than relying on prescriptions, they used e-prescribing administration records — the nurse's confirmation that a dose was given — and paired those with routinely collected vital signs.

An event counted as respiratory depression if a patient's breathing fell below 10 breaths per minute; below 12 with oxygen saturation under 94 percent; below 12 with reduced consciousness; or if naloxone was administered to reverse an overdose. Just under 2,000 patients, about 6 percent of the cohort, met that definition. Naloxone was given to 318.

"A key strength of our study was our ability to use detailed hospital electronic health records to accurately capture when opioids were actually administered to patients, alongside routinely collected vital signs to identify changes in breathing," said Dr. Meghna Jani, an NIHR Advanced Fellow and senior clinical lecturer at Manchester who was the study's corresponding author.

The rate of respiratory depression while patients were on an opioid was 10.20 events per 1,000 person-days, against 2.41 per 1,000 when they were off it.

Where each drug landed

Using codeine as the comparison drug and adjusting for age, sex, ethnicity, comorbidity burden, deprivation, kidney disease, sleep apnea, and alcohol use, the risks separated sharply:

Fentanyl carried a hazard ratio of 3.36. Combination opioid products came next at 2.74, followed by oxycodone at 2.10 and morphine at 1.84. Topical buprenorphine patches (1.74) and tramadol (0.87) did not reach statistical significance, and the authors caution that too few patients received them to estimate the risk precisely.

Measured against morphine instead, fentanyl still ran 85 percent higher. Combination products were 49 percent higher, while codeine sat at roughly half morphine's risk.

One point deserves care, because early coverage of the study garbled it: codeine was the reference category, not a demonstrated safest option. Against morphine, tramadol scored lower than codeine did, and the residual group of less-common opioids lower still. The study establishes a gradient among potent opioids; it does not crown a winner at the bottom.

Patients with chronic obstructive pulmonary disease fared worse across the board. In that subgroup, fentanyl's hazard ratio climbed to 4.03.

The dose thresholds start lower than guidelines assume

Risk rose with dose in a curve rather than a straight line, and it began rising earlier than current prescribing advice implies. Patients receiving 120 or more morphine milligram equivalents per day had roughly double the risk of those under 50. But the increase was already measurable in the 31-to-60 range, a band most guidance treats as routine.

The authors note that the British Pain Society recommends a ceiling of 120 MME per day for chronic pain, and write plainly that "our findings support lowering it." They also argue clinicians should "avoid fentanyl as a first-line potent opioid" and that non-opioid and non-drug approaches "should be prioritised whenever possible."

Adding a gabapentinoid nearly doubled the danger

The sharpest practical finding concerns what else patients were taking. Receiving an opioid alongside a gabapentinoid — gabapentin or pregabalin, commonly prescribed for nerve pain, anxiety, and sleep — carried a 73 percent higher risk of respiratory depression than the opioid alone, and nearly four times the risk of patients on neither.

"The use of gabapentinoids with opioids in particular was associated with an increased risk of respiratory depression," said Carlos Raul Ramirez Medina, a research associate at Manchester and the study's first author.

That signal aligns with US mortality data. A CDC analysis reported earlier this year found gabapentin detected in a rising share of overdose deaths, with the large majority of those cases also involving opioids. The Manchester study captures the same interaction upstream, in the ward, before it becomes a death certificate.

Benzodiazepines produced a counterintuitive result: patients on an opioid plus a benzodiazepine showed lower risk. The authors attribute this to channelling bias — the combination is already flagged as dangerous, so prescribers reserve it for patients they judge to be at low risk and monitor more closely. It is a finding about physician behavior, not pharmacology.

What the study cannot tell you

Several limits matter, and some of them were lost in the first wave of coverage.

The study has no mortality endpoint. It measured respiratory depression events, not deaths, and cannot rank opioids by how likely they are to kill anyone. Headlines framing it as identifying the deadliest painkillers overstate what was tested.

It is also a study of therapeutic use. These were hospital inpatients receiving opioids from nurses for non-cancer pain. Patients on methadone or sublingual buprenorphine were deliberately excluded because those medications signal opioid substitution therapy — meaning the findings say nothing about the safety of medication-assisted treatment for opioid use disorder, and nothing about illicit fentanyl, which reaches the bloodstream in uncontrolled doses under entirely different circumstances.

Finally, this is one tertiary hospital in one English region, with data now roughly nine years old. The authors acknowledge that vital sign recording may have been inconsistent, that they could not adjust for frailty or low body mass index, and that estimates for the less-frequently used opioids rest on small numbers.

Why prescribing safety is an addiction story

The relevance to opioid use disorder is indirect but real. Hospital exposure is where a substantial share of patients first encounter potent opioids, and the drug chosen there often follows them home on a discharge prescription. A finding that fentanyl and combination products carry distinctly higher physiological risk, at doses lower than guidelines flag, argues for choosing the least potent effective agent — the same principle that limits how many people move from prescribed opioids to dependence.

The gabapentinoid result carries the most immediate weight. Co-prescribing is common, largely unmonitored, and now associated with a near-doubling of respiratory risk in a population that was not otherwise considered high-risk.

"Opioids remain important medicines for managing severe acute pain," Jani said. "Our findings show that the risks are not the same across all opioid drugs or doses." Understanding those differences, she added, can help clinicians and patients "make more informed prescribing decisions together, as well as increasing awareness of what dose thresholds require closer monitoring."