Is long-term paracetamol use not as safe as we thought?

Tuesday March 3 2015

"Daily paracetamol could raise the risk of heart attacks, stroke and early death," the Mail Online reports.

A new review of previous observational studies found that long-term use of paracetamol was linked with a small increased risk of adverse events such as heart attacks, gastrointestinal bleeds (bleeding inside the digestive system) and impaired kidney function.

It is important to be aware that, as these are observational studies, there is the potential for various sources of bias. The studies were highly variable in their study populations. For example, four studies included female nurses, one male doctor, one person with kidney disease, and other adults prescribed paracetamol (i.e. they weren’t taking it over the counter). They also examined highly variable paracetamol exposures (e.g. days of use per month, grams intake in a lifetime, or number of prescriptions). Overall, this gives quite a mixed group of study designs and results, which may be influenced by many things, including inaccurate estimations of intake and significant health differences between users and non-users of paracetamol.

Nevertheless, the findings that paracetamol could potentially have adverse longer-term effects, particularly when used at higher doses, is important, especially as the drug is used by millions. Therefore, further investigation is needed.

Where did the story come from?

The study was carried out by researchers from the Maudsley Hospital, London; University of Leeds; Newcastle University; Keele University; and other UK institutions. The review was undertaken by the National Clinical Guidelines Centre, UK, and the authors report no conflict of interest.

The study was published in the peer-reviewed Annals of the Rheumatic Diseases, a British Medical Journal, on an open-access basis, so it is free to read online or download as a PDF.

Some of the UK’s media reporting is likely to cause needless alarm, particularly with figures such as the Mail’s "63% more likely to die suddenly" and claims that "risk of heart attack or stroke 68% higher". Such reporting takes a rather simplistic view of research that has included a highly variable mix of studies.

Ideally, the studies that have contributed different risk figures would benefit from being considered on an individual basis (i.e. taking into account their specific study designs and methods, and potential biases) rather than it simply being applied to the general population.

For example, the 63% figure for mortality accounted for adults prescribed paracetamol or ibuprofen in one particular study participant who had been receiving the highest number of repeat prescriptions for paracetamol, with the shortest gaps between prescriptions. The 68% risk figure for heart attacks was for US female nurses taking more than 15 paracetamol tablets a week.

What kind of research was this?

This was a systematic review of observational studies that aimed to look at the adverse effects of paracetamol.

As the researchers say, paracetamol is the most widely used over-the-counter and prescription painkiller worldwide, and is often the first painkiller taken for a wide variety of conditions. It is generally considered to be safer than other painkillers that may be considered in later steps on the "pain ladder", such as non-steroidal anti-inflammatory drugs (NSAIDs) or opiates. However, both health professionals and patients need to have up-to-date evidence on the possible harms of a drug, and recent estimates of the possible risks of paracetamol are not currently available. Therefore, this review aimed to address this gap.

A systematic review is the best way of gathering all the available studies that have addressed the effects of a particular treatment. However, the findings of the review are always going to be inherently limited by the underlying studies. The best way of looking at the benefits and harms of a treatment is a randomised controlled trial. However, it is not ethical to randomise a person to take, for example, a daily dose of paracetamol for a long period of time purely to look at its adverse effects.

When looking at adverse effects of a treatment in observational studies, there is always the possibility that results are being influenced by other factors, such as health differences between people who choose to take the treatment or not. This is known as channelling bias – people who are in poor health are more likely to be "channelled" on to a specific drug regime than people who are healthy.

What did the research involve?

The researchers searched two literature databases up to May 2013 to identify observational studies in adults (aged >18 years) that had examined the adverse effects of taking standard dose oral paracetamol (0.5 to 1g, 4 to 6 hourly to a maximum of 4g per day) compared with non-use.

The main outcomes examined were all-cause mortality, cardiovascular effects (specifically heart attacks, strokes and high blood pressure), gastrointestinal (gut) effects (specifically bleeding), and kidney effects (poorer kidney function as indicated by kidney filtration rate, blood chemistry or need for replacement therapy, such as dialysis).

What were the basic results?

Eight studies met the inclusion criteria, all were cohort studies. Five were conducted in the US, one in the UK, one in Sweden and one in Denmark. Included sample size ranged from 801 to 382,404 participants, and duration of follow-up ranged between two and 20 years. The studies included specific populations:

  • four of the US studies included female registered nurses aged 30-55 years
  • the other US study included male doctors
  • the UK study involved people over 18 years prescribed paracetamol or ibuprofen
  • the Swedish study included people diagnosed with chronic kidney disease
  • the Danish study comprised of people over 16 years (though the systematic review inclusion criteria did specify 18 years) that were prescribed paracetamol

The studies each looked at some of the outcomes being studied, and examined various levels of exposure compared with non-use. For example, some looked at number of days use per month (e.g. one to four days, ranging to more than 22 days); others looked at number of grams (g) of lifetime intake (e.g. ranging from 100g lifetime intake to >3000g); the gaps between prescriptions; another looked at the number of pills taken over a set 14-year time period.

Looking at outcomes:

  • One of two studies looking at mortality reported an increased mortality ratio for paracetamol users compared with non-users (overall 28% increased risk). This was in a study of adults prescribed paracetamol or ibuprofen. Sub-analyses found the highest risk with the greatest number of prescriptions (63% increased risk).
  • Four studies found paracetamol use was associated with cardiovascular effects, with increased risk linked to increased exposure. One study found a 68% risk of cardiovascular events for people (one of the nurses’ studies) who took more than 15 tablets per week. Another study also found higher doses was associated with heart attack and stroke, and two others found associations with high blood pressure.
  • One study reported gastrointestinal effects and found overall (36%) increased risk of gastrointestinal bleeds. This was in the study of adults prescribed paracetamol or ibuprofen, with the highest associated with first prescription (74%) and greatest number of prescriptions (49%).
  • Four studies reported kidney effects, and three found a poorer kidney function with increasing dose.

How did the researchers interpret the results?

The researchers conclude that, "the dose-response seen for most endpoints suggests a considerable degree of paracetamol toxicity, especially at the upper end of standard analgesic doses". However, they give due caution that, "given the observational nature of the data, bias may have had an important impact".


This is a valuable review that has searched the literature and identified eight observational studies in adults that have looked at the adverse effects that may be associated with paracetamol use. As the researchers say, it will add to information on the potential harms of paracetamol – an area where up-to-date information has been lacking.

The studies included very large population sizes, and collectively provide some evidence suggesting potential effects upon the cardiovascular system, kidney and gastrointestinal system. There was also a suggestion of increased risk of all-cause mortality.

However, it is very important that these findings are interpreted in the right context. The eight studies were a very variable mix, with very different populations, study duration, measurement of paracetamol exposure and outcomes examined. Most of the individual risk figures identified and reported in the media (e.g. 63% increased risk of mortality or 68% increased risk of heart attack) actually came from individual studies, and would ideally benefit from being interpreted in the specific context of that study. For example, the 68% risk figure for heart attacks was for US female nurses taking more than 15 paracetamol tablets a week.

A possibility for all of these studies is that their results could be influenced by confounding factors (confounders); that is, other differences between the non-users and users of paracetamol that are accounting for the differences. When looking at people using the greatest amounts of paracetamol in these studies – those with, generally, the highest risk figures – their health differences compared to non-users may be even more considerable. For example, increased paracetamol use was associated with increased risk of gastrointestinal bleeds, but this was in a study where people were prescribed paracetamol or ibuprofen.

Ibuprofen, like other NSAIDs, is known to be associated with gastrointestinal bleeds, and it could be that people needing to take more paracetamol were also needing to take more ibuprofen, which could have been influencing the risk. Similarly, other studies included people who may have been suffering from various chronic diseases, which could have increased their risk of cardiovascular events, poorer kidney function and mortality, and also cause them to need more painkilling medication.

It is not known whether, or how well, these studies took into account all the potential health and lifestyle differences that may be associated with increased paracetamol use, as well as increased risk of adverse health outcomes.

Another potential source of inaccuracy is in estimations of paracetamol intake. For example, it is difficult to know how someone could reliably estimate the number of grams of paracetamol they had taken in a lifetime, or how many pills they had taken over a 14-year period.

Overall, the findings that paracetamol could potentially have adverse longer-term effects, particularly higher doses, are undoubtedly important and will need further investigation. 

Paracetamol is an effective treatment for mild to moderate pain and fever in adults and children, when used as directed in product information. The maximum dose within a 24-hour period must not be exceeded. However, if you find you need to use paracetamol on a regular basis, it is worth consulting your GP to look at the cause, and possible treatments. You may find your symptoms respond better to an alternative painkiller or possibly a non-drug type of treatment, such as physiotherapy.

There are also a number of self-help techniques that can help people cope better with chronic pain.

Analysis by Bazian
Edited by NHS Choices