At the tail end of August, Professor Trond Riise and his team at the University of Bergen’s Department of Global Public Health and Primary Care announced an important discovery. Riise’s Norwegian team, in collaboration with researchers from Harvard University, have confirmed a significant link between the use of certain medications and their users’ risk of developing Parkinson’s disease (PD).
The study, published in September in Science, found that using the asthma medication salbutamol halved the risk of Parkinson’s, while propranolol, a beta blocker often prescribed for high blood pressure, was found to have the opposite effect, virtually doubling Parkinson’s risk.
The mechanism explaining these risk profiles is the beta-2 adrenergic receptor, which is a regulator of the alpha-synuclein (SNCA) gene that is abundant in the brain. Excess production of SNCA is a potential causative factor in the development of PD, so the researchers at Harvard suspected that beta-2 agonist drugs could reduce PD risk, while beta-2 antagonists could increase the risk. The American researchers had observed this effect in genetic and animal studies, but Riise’s team in Norway, which looked for the same link in human data, was surprised that its results matched animal models so closely.
“We found exactly the same thing as they found in the mice in the US,” says Riise. “We said, ‘Wow, this is too good to be true.’ But it turned out to be correct.”
The validation of this hypothesis in human data is an important discovery that helps inform our understanding of PD, and the risks associated with drugs that suppress the beta-2 receptor. It could also kick off a new research avenue looking into new ways to treat and prevent the disease. And none of it would have been possible without the key piece of the puzzle – comprehensive prescription data taken from the Norwegian Prescription Database (NorPD).
NorPD: vital to the Parkinson’s risk discovery
NorPD is the reason that the American researchers came to Norway for help with the massive quantity of data necessary to validate the link to PD risk. Like the vast majority of countries around the world, the US doesn’t have the comprehensive data to track the health of people taking specific prescription drugs out of the whole population. Norway, on the other hand, had exactly this capability with NorPD, a nationwide ledger of every prescription issued to the Norwegian public since 2004, anonymised to protect the identities of individuals, but including data-points such as the age and gender of those taking each prescription.
Starting in 2012, Riise and his colleagues at the University of Bergen spent several years delving into this massive database. By the time they were ready to confirm the link, they had analysed more than 100 million prescriptions ordered over the previous 11 years. The team identified Norwegians who had developed Parkinson’s from those who had been prescribed levodopa, the key first-line drug for management of PD, and who had been taking the drug for at least a year. Cross-referencing with the asthma and hypertension drugs in question, not to mention separating the data into individuals rather than simply prescriptions, was no mean feat.
“We accumulated, for each individual, a number of daily defined doses for each of the medications, per year,” Riise says. “We had the various syntaxes to get down to five million people. For each of these five million people, for each year, we had the amount of drugs being prescribed for these individuals. So that was the first round, to reduce it from separate prescriptions down to individuals, and then sum up all the prescriptions for each individual.
“That took time – we would start up the computers and they went and went for hours to get down to these numbers. And then we ran pretty good statistical models that looked at the risk of developing the disease based on exposure that could vary.”
For the most part, Riise’s experience of using data from NorPD has been extremely positive; he praises the database for its comprehensiveness and reliability. “The beautiful thing with NorPD is that there are no mistakes,” he says. “All the prescriptions are filed in a way – it’s not manual, it’s automatic based on the codes. So that’s a very nice thing; there’s no missing data and no errors in the data.”
Nordic leadership: the development of NorPD
Norway isn’t entirely alone in having comprehensive data on population-wide prescription drug use. The four other Nordic countries – Sweden, Denmark, Finland and Iceland – have also established this capability, and so too have South Korea and Taiwan. Nevertheless, NorPD is a rare asset; since the database was first launched in 2004, there have been dozens of epidemiological studies published using the information, analysing topics from prescription opioid misuse to drug adherence. So how did this valuable database develop, and why has Norway taken such a leading position on prescription data tracking at this scale?
According to Dr Kari Furu, a senior researcher at the Norwegian Institute of Public Health’s Department of Pharmacoepidemiology (who worked for three years at the Ministry of Health in the early 2000s to help establish NorPD), Norway has a long history of wide-ranging drug statistics, having been the first country in the world to produce public drug statistics in the 1960s. Nevertheless, the roots of NorPD lie in the mid-90s.
“We became a member of the European Economic Association at that time,” says Furu. “After ’95, relatively more new medicines were being introduced per year in the Norwegian market. So the discussion started in about ’95. The main argument from the government was to see the drug utilisation in the population as a public health issue, actually, because when you have the wholesale statistics, you can’t see anything about who is using [the drugs] – age, gender and so on. You can’t do safety studies, and you can’t see changes in the market after marketing new drugs.”
A key string to NorPD’s bow is its focus on research. The Department of Pharmacoepidemiology was established alongside the database, with the new research unit serving a dual role – to hit the ground running with research based on the data, and to make NorPD more accessible to those who weren’t familiar.
“Without the researchers linked to this registry, the data would not have been used as much as it has been,” says Furu. “We travelled around and gave presentations for other research environments and for people in the health ministry and other health authorities, because it’s quite complex and some researchers called it messy. It’s very good quality data, but there are some challenges to use these huge databases.”
Broader, deeper, faster: improving NorPD for researchers
Establishing the link between certain drugs and Parkinson’s risk is a recent highlight of the studies that have come out of the NorPD endeavour, but it’s by no means alone. Other high-impact research that Furu has taken part in includes a number of studies on the impact of muscle relaxant drug carisoprodol, which, the studies demonstrated, was being misused recreationally by some users. As a result of these insights, carisoprodol was removed from European markets in 2007.
As well as providing warnings on potential unexpected risks presented by drugs on the market, NorPD data has also been used to provide reassurance that a much-feared drug health risk doesn’t exist.
“I have been especially working with the prescription database linked to the medical birth registries to analyse outcomes in children where the mother has been using [anti-depressants] during pregnancy,” Furu says. “In 2015, I was the first author on an article in the BMJ [British Medical Journal], and we could assure pregnant mothers that it’s not a huge increase in the risk of malformations when using anti-depressants during pregnancy. We are also doing studies to reassure pregnant women because they are afraid of doing something wrong.”
Nevertheless, there are ways in which NorPD could be improved for the future. The database currently does not cover prescriptions issued to in-patients in hospitals and nursing homes, and Furu would like to see this gap closed.
The time it takes for researchers to get access to NorPD’s data could also be improved, especially when they want to link the database to one or more of the Nordic countries’ many other public health registers.
“It’s still cumbersome and it takes time,” says Furu. “Sometimes you have to wait a year to link [between databases and registers], and we have drug safety questions that we want to answer very quickly.”
“The data are available and it’s a very good source, but it’s not incredibly easy to have all kinds of approval, because of the concern of confidentiality,” says Riise. “So you have to do some major work.”
Nevertheless, the upsides significantly outweigh the drawbacks for researchers making use of this treasure trove of prescriptions data. The prospect that the next big discovery is hiding deep within the database, waiting to be discovered, is a tantalising one. Riise says there are likely more clues in the NorPD database to further elucidate the Parkinson’s mechanism that has been discovered, to name but one avenue for exploration. The benefits of having comprehensive national prescription data seem to be many-fold; the many countries that have not followed the Nordic nations’ lead could do well to start catching up now.