Thursday March 3 2011
Prolonged ibuprofen use carries risks and should be avoided
Taking ibuprofen just a few times a week could cut the risk of developing Parkinson's disease by a third, according to The Daily Telegraph.
The news is based on the publication of a large study that followed 136,197 middle-aged to elderly people over six years. It looked at whether regular use of the painkiller ibuprofen had any association with the risk of developing Parkinson’s disease. The study found that 291 people developed Parkinson’s, with those who regularly took ibuprofen having an approximately 30% lower risk of developing the disorder than those who did not. Other painkillers were also examined but were not associated with a reduced risk.
The study was well-designed but has some limitations, which mean it cannot prove that ibuprofen can help to protect against Parkinson’s. For example, only 28 people who developed Parkinson’s had used ibuprofen, making it hard to perform statistical comparisons of their behaviour. Also, early (pre-clinical) Parkinson’s may be present many years before obvious symptoms, so it is possible that participants may have already had undiagnosed Parkinson’s before their ibuprofen use was assessed.
Regular use of ibuprofen can have side effects, including a raised risk of stomach bleeds. Given the potential risks, people should not attempt to take ibuprofen as a preventative treatment against Parkinson’s disease at the current time.
Where did the story come from?
This US study was carried out by researchers from Brigham and Women’s Hospital, Harvard medical School, Harvard University School of Public Health, the National Institute of Environmental Health Sciences, and Massachusetts General Hospital. The research paper featured no information about external funding. The study was published in the peer-reviewed medical journal, Neurology.
In general, the study was reported accurately by the newspapers, although reports tended to be over-optimistic and did not mention the study’s limitations.
What kind of research was this?
This research was based on data from two prospective cohort studies, and it involved more than 136,000 participants. It looked at whether use of ibuprofen, other NSAIDs or paracetamol was associated with a lower risk of developing Parkinson’s disease: a progressive neurological disorder characterised by muscle tremor, stiffness and weakness.
This type of study, which can follow large groups of people for many years, is useful to assess the possible relationship between an intervention (in this case, use of ibuprofen and other painkillers) and an outcome (in this case, development of Parkinson’s disease). However, on its own it cannot prove a causal association between the two. Prospective cohort studies, which follow people in real time, are also considered to be more reliable than retrospective studies, which often ask people to recall events that have happened several years in the past.
The authors also pooled the results of their study with other previously-published trials to perform a meta-analysis of the relationship between NSAIDS, other painkillers and Parkinson’s.
The researchers discuss how neuroinflammation, a chronic, inflammation-like response in the central nervous system) may contribute to the development of Parkinson’s Disease. They point out that previous epidemiological studies suggest that use of NSAIDs in general, and ibuprofen in particular, may be related to a lower risk of developing Parkinson’s.
What did the research involve?
The researchers used data from two very large, long-term studies of health professionals. One was based in the US (the Health Professionals Follow-up Study, which began in 1986) and one was from the UK (the Nurses’ Health Study, which began in 1976). Both studies are based on participants completing questionnaires regarding the medical history and lifestyle of participants at the start of each study, with follow-up questionnaires mailed every two years.
The authors have already published earlier research from these groups, which found an association between non-aspirin NSAID use and a lower risk of PD. This new research was restricted to the years after the original study, using the 2000 US survey and the 1998 UK survey as their starting point. The total number of participants in these studies was 136,197.
The researchers established that participants were not diagnosed with Parkinson’s at the start of their study. They assessed the use of NSAIDs by questionnaire, with participants being asked if they regularly took (two or more times weekly) the painkillers aspirin, ibuprofen, other NSAIDs or paracetamol. Information on the participants’ use of these painkillers was updated every two years for both study groups. The questionnaires also recorded information on age, ethnicity, body weight, height and smoking status.
Participants were followed for six years. Those diagnosed with Parkinson’s over this period were identified using self-reports and diagnostic confirmation from relevant doctors.
The researchers used standard statistical techniques to assess the possible relationship between use of NSAIDs and Parkinson’s. They adjusted their findings to take account of possible “confounders” which might have affected results, including age, smoking and caffeine intake. Researchers also excluded patients with gout, since high uric acid levels also lower PD risk. They excluded PD cases identified in the first two years of follow-up, to avoid the possibility of reverse causation, i.e. people not taking NSAIDs because of their PD.
What were the basic results?
During six years of follow-up, the researchers identified 291 people who had developed PD. They found that:
- After adjusting for age, smoking, caffeine use and other possible confounders, people using ibuprofen had a significantly lower PD risk than non-users (relative risk [RR], 0.62, 95% confidence interval [CI] 0.42 to 0.93).
- The higher the dose of ibuprofen taken each week, the lower the risk. This is called a dose–response relationship.
- Use of other painkillers, including aspirin, paracetamol and other NSAIDs, had no significant association with risk of PD.
- When researchers combined their results with other published studies in a meta-analysis, they again saw a reduced incidence of Parkinson’s disease with ibuprofen use (pooled RR of Parkinson’s 0.73, 95% CI 63 to 0.85).
- In the meta-analysis, other types of analgesics were once again not found to be associated with lower Parkinson’s disease risk.
How did the researchers interpret the results?
The researchers say their results suggest that ibuprofen should be investigated further as a “potential neuroprotective agent” against Parkinson’s disease. They add that there is some evidence that “inflammatory mechanisms” may contribute to the progressive damage to nerve cells. They argue that ibuprofen therefore possibly has protective properties against this process. They suggest that these protective properties are not shared by other NSAIDs.
The strengths of this study lie in its large sample size and high follow-up rate (95% and 94% in the UK and US studies, respectively). Because the study was prospective, following people in real time, there was also less chance of “recall bias” (where participants inaccurately recall the use of painkillers). In addition, the researchers controlled for important confounding factors, such as age, smoking, body mass index, caffeine and alcohol intake. The way they assessed NSAID use, intended to cover both prescription and over-the-counter use, is also thought to be reliable.
However, as the authors note, it does have some limitations:
- NSAID use was self-reported and therefore potentially subject to error.
- The studies involved US and UK health professionals rather than random samples of men and women. Their use of NSAIDs would not necessarily reflect the pattern of use seen in the general population. The authors point out that the biological effects of ibuprofen on Parkinson’s disease would be the same, however.
- It is possible that ibuprofen was used to treat conditions that are themselves associated with a lower risk of PD. That said, the primary use of ibuprofen was for muscle and joint pain, which is not associated with PD risk.
- Although they adjusted for confounders, other factors that might influence the results cannot be ruled out.
Importantly, although this was a large study, it should be noted that the number of people who developed Parkinson’s disease was small (28 Ibuprofen users and 263 non-users). Carrying out statistical comparisons in such few participants can be problematic as it increases the possibility of presenting inaccurate risk associations. The potential for inaccuracy is even greater when subdividing them by dose taken. For example, only nine people with Parkinson’s had taken ibuprofen once or twice a week; four people used it three to five times a week; and 10 people, more than six times. Though they observed a trend for higher dose to be associated with lower risk, this may therefore be inaccurate.
A further limitation that might have affected results is the study’s short follow-up period: as an accompanying editorial points out, early signs of “preclinical” PD may be present up to 20 years before obvious symptoms. It is possible that gastrointestinal symptoms, for example, could cause a person with very early Parkinson’s to be less likely to take ibuprofen regularly (because it would be contraindicated).
In conclusion, this study is of interest but it cannot show a causative association between ibuprofen use and the development of Parkinson’s. Further research is required to investigate whether ibuprofen could be “neuroprotective”.
Regular use of ibuprofen and other NSAIDs can have side effects, including stomach bleeds, particularly in the elderly, and a slightly increased risk of heart attack and stroke. Given these risks, and the uncertainy over whether it is associated with a lower risk of Parkinson’s disease the use of ibuprofen as a preventative treatment against Parkinson’s cannot be recommended at the current time.