In an effort to better understand Parkinson’s, researchers have repeatedly analysed data from large epidemiological studies in order to gain insight into factors that could have a possible causal influence in the development of the condition.
This week a manuscript was made available on the preprint website BioRxiv that provided us with a large database of information about aspects of life that are associated with increased incidence of Parkinson’s.
Some new associations have been made… and some of them are intriguing, while others are simply baffling!
In today’s post, we will have a look at what has been learnt from epidemiological research on Parkinson’s, and then discuss the new research and what it could mean for Parkinson’s.
What are the differentiators? Source: Umweltbundesamt
What makes me different from you?
Other than my ridiculous height and the freakishly good looks, that is. What influential factors have resulted in the two of us being so different?
Yes, there is the genetics component playing a role, sure. 7,500 generations of homo sapien has resulted in a fair bit of genetic variation across the species (think red hair vs brown hair, dark skin vs light skin, tall Scandinavians vs African pygmies, etc). And then there are aspects like developmental noise and epigenetics (factors that cause modifications in gene activity rather than altering the genetic code itself).
And over-riding all of this, is a bunch of other stuff that we generally refer to simply as ‘life’. Habits and routines, likes and dislikes, war and famine, etc. The products of how we interact with the environment, and how it interacts with us.
But which of all these factors plays a role in determining our ultimate outcome?
It is a fascinating question. One that absorbs a large area of medical research, particularly with regards to factors that could be influential in causing a specific chronic conditions.
What does this have to do with Parkinson’s?
Well, researchers have been very interested in determining which factors could possibly differentiate people who are going to develop Parkinson’s from those who will not? That is to say, which aspects of life (beyond genetics) are associated with the risk of developing Parkinson’s?
And there have been some epidemiological studies have already pointed towards certain factors that appear to be having ‘an influence’.
What are epidemiological studies?
Epidemiology is the study of the distribution (the who, when, and where) and the determinants of health or disease within defined populations. Basically, it is the process of trying to work out why certain people get sick while others do not.
Your typical epidemiological study takes a large group of people and follows them over time, recording significant events and outcomes.
And some of these epidemiological studies have taught us a great deal about Parkinson’s.
Let’s look at just one example – it’s a classic:
In 1950, Dr Travis Gordon noticed something very curious.
While the overall mortality rates for men in the USA and Japan were very similar, the incidence of heart disease was significantly lower in Japan. This single observation resulted in three large longitudinal studies – one of which became known as the Honolulu Heart Study.
Dr Travis Gordon. Source: JSTOR
The original purpose of the Honolulu Heart Study was to simply determine whether there was a difference in the incidence of heart disease between Japanese people living in Japan and individuals of Japanese ancestry living in the US state of Hawaii.
That was the original goal, but understand that this study has provided soooo much more that just that.
The participants were recruited for the study were “non-institutionalised men of Japanese ancestry, born 1900-1919, resident on the island of Oahu.” In all, 12,417 men were identified as meeting the criteria. Of those contacted, 1,269 questionnaires were ‘return to sender’, 2,962 men declined to participate in the study, and 180 died before the study commenced. That left 8,006 participants who would be studied and followed for the rest of their lives.
Honolulu – a tough place to stay focused on research. Source: Hiltonhawaiianvillage
From October 1965 onwards, the participants were interviewed and given physical examinations every few years. The interview processed asked for:
- Family and personal history of illness
- Sociological history
- Smoking status
- Physical activity level
The physical examination was very thorough, looking at:
- ECG (Electrocardiography – electrical activity of the heart)
- Urine analysis
- Measurements of weight, height, skinfold thickness, etc.
- Blood pressure and serum cholesterol
As a result of this massive data collection effort, the Honolulu Heart Study has built up a HUGE amount of epidemiological information regarding these 8,006 individuals.
But what does this have to do with Parkinson’s????
Given the enormous number of individuals involved in the study and the length of time that they were followed, it was inevitable that a certain percentage of them would develop Parkinson’s as the study progressed. And given this situation, for a long time the Honolulu Heart Study represented one of the largest epidemiological study of Parkinson’s.
In 1996, a group of researchers involved in the study, published some very interesting findings relating to Parkinson’s:
Title: Epidemiologic observations on Parkinson’s disease: incidence and mortality in a prospective study of middle-aged men.
Authors: Morens DM, Davis JW, Grandinetti A, Ross GW, Popper JS, White LR.
Journal: Neurology, 1996 Apr;46(4):1044-50.
In total, 92 of the 8006 individuals enrolled in the study had developed Parkinson’s. The incidence of Parkinson’s cases was registered between 1965 and November 30th 1994. The majority of the cases were diagnosed between 55 and 79 years of age (n=80). Diagnosis after the age of 80 was very rare. It is interesting to note that when the researchers divided the group into those ‘born before 1910’ and those ‘born after 1910’, the older group (born before 1910) actually had a lower incidence of Parkinson’s. In addition, the incidence rate in the Hawaiian cohort was similar to rates previously published for Caucasian men in Europe and the US, which made it higher than the previously published incidence rates for Asian men living in Asian nations. The nation of Japan (as a population) has a lower incidence of Parkinson’s (much like most of the East Asian nations – click here to read more on this) than their western equivalents, despite actually living longer. Thus, the finding that Japanese individuals living in Hawaii have a similar incidence of Parkinson’s to western nations is very interesting and suggestive of environmental influences.
ANOTHE INTERESTING FACT ABOUT THE JAPANESE: Japan is the only country in the world where females have a higher incidence of Parkinson’s than men (Click here to learn more about this).
In another research report, the same group of Honolulu Heart Study investigators made even more interesting discoveries:
Title: Prospective study of cigarette smoking and the risk of developing idiopathic Parkinson’s disease.
Authors: Grandinetti A, Morens DM, Reed D, MacEachern D.
Journal: American Journal of Epidemiology 1994 Jun 15;139(12):1129-38.
In this report, the investigators found that men who had smoked cigarettes at any time prior to their enrollment in the study in 1965, had a reduced risk of developing idiopathic Parkinson’s. That is to say, smoking reduced the chance of developing Parkinson’s. The relative risk was 0.39, and when we talk about relative risk, anything below 1.0 indicates a negative association (or a reduced risk of a particular condition), while anything >1.0 indicating a positive association (or an increased risk of a particular condition).
A few years later the authors published a follow-up paper which rejected the possibility that smoking was killing people before they could develop Parkinson’s (or the possibility that selective mortality results in a false positive). That follow-up report can be found here.
EDITOR’S NOTE: These results do not mean you should rush out and start smoking. Smoking causes other conditions that can affect your health and result in death.
As you can see in the table below, this smoking finding was made alongside other interesting correlations (Note: coffee and alcohol reduce the risk of Parkinson’s):
From Grandinetti et al (1994).
It should be noted that many of these associations (smoking in particular) had been reported anecdotally before, but the Honolulu Heart Study was the first epidemiological study to provide ‘definitive’ proof. And it should also be noted that subsequent epidemiological studies have found similar results (Click here to read more about this).
Interesting. So what is the new research released this week?
This is Anne Wojcicki:
Anne Wojcicki Source: Speakerpedia
She is cool.
She is also the CEO and Co-Founder of 23andMe – a personalised genomics company. 23andMe can tell you a great deal about yourself by analysing your DNA (for examples, where your ancestors are from and certain medical conditions you could be predisposed to).
23andMe have long had a strong association with the Parkinson’s community (Click here to read more about this), and that association has recently been paying big dividends.
Last year, for example, the company published this research report based on data that they have been collecting since 2009:
Title: A meta-analysis of genome-wide association studies identifies 17 new Parkinson’s disease risk loci.
Authors: Chang D, Nalls MA, Hallgrímsdóttir IB, Hunkapiller J, van der Brug M, Cai F; International Parkinson’s Disease Genomics Consortium; 23andMe Research Team, Kerchner GA, Ayalon G, Bingol B, Sheng M, Hinds D, Behrens TW, Singleton AB, Bhangale TR, Graham RR.
Journal: Nature Genet. 2017 Sep 11. doi: 10.1038/ng.3955.
In this study, the researchers reported the discovery of 17 new regions of DNA that are are associated with increased risk of developing Parkinson’s. Most research reports highlight just one or two novel genetic risk factors – this report gave us 17!!! And the analysis was based on a HUGE dataset (DNA from 26,035 cases of Parkinson’s and 403,190 controls samples). We will have a more thorough review of this research in an up-coming post revisiting the genetics of Parkinson’s.
But moving on to other dividends stemming from the 23andMe/Parkinson’s relationship: this week the company released new research looking at a non-genetic aspect of Parkinson’s.Researchers at 23andMe in collaboration with the good folks at the PredictPD study made a manuscript available on the preprint website BioRxiv – a website that provides access to manuscripts before they have been peer-reviewed in order to speed the rate of information sharing.
This is the manuscript in question:
Title: The Parkinson’s Phenome: Traits Associated with Parkinson’s Disease in a Large and Deeply Phenotyped Cohort
Authors: Heilbron K, Noyce A, Fontanillas P, Alipanahi B, The 23andMe Research Team, Nalls M, Cannon P
Journal: bioRxiv preprint first posted online 28th February, 2018
PMID: N/A (This article is OPEN ACCESS if you would like to read it)
In this study, with help from the Michael J. Fox Foundation, the researchers recruited 5,141 people with Parkinson’s and 65,459 age- and sex- matched controls between 2007 and 2017 and asked them to answer questions about 840 phenotypes using web or mobile-based surveys .
What is a phenotype?
Phenotypes are observable characteristic produced by an individual that results from the interaction of its genotype (or it’s genetic constitution) and the environment. Phenotypes include things like morphology, biochemical or physiological properties, and critically behaviours (everything from hobbies and social skills to sleep and exercise patterns). Phenotypes are all of the observable characteristics of an organism as it exists and interacts with a particular environment.
The investigators found that a diagnosis of Parkinson’s was associated with 149 independent phenotypes.
They found 32 of those phenotypes were known associations (such as REM sleep behaviour disorder having a positive association with Parkinson’s, while drinking coffee and smoking had a negative association). But they also discovered 49 associations that have never been previously reported (see the list below).
Interestingly, they found that migraines, obsessive-compulsive disorders, seasonal allergies, and anemia (a reduction in red blood cells) were associated with Parkinson’s, but all of these tended to have an age of onset several decades before the average age of Parkinson’s diagnosis.
In the figure below is the list of the 149 phenotypes that were significantly associated with Parkinson’s. The dashed line down the centre indicates no association (or 1.0). Circles leaning towards 0.0 have a negative association (or a reduced incidence with Parkinson’s), and circles leaning towards 2.0 have a positive association (or increased incidence with Parkinson’s):
Lots of well known associations in these figures (such as constipation and restless leg syndrome). But some of the novel associations are rather…well,…bewildering.
For example, under negative associations (reduced incidence of Parkinson’s):
- Learning a language as an adult
- Uses earplugs to sleep (I love the granularity of some of these phenotypes!)
- Could do forward splits
- Adopted (wow!)
- Bothered by others chewing
- Can wiggle ears
- Ring finger: index finger ratio
- Clap to a beat
- Can whistle
- Cosmetic surgery
- Liposuction (?!?)
And under positive associations (increased incidence of Parkinson’s):
- Fear of heights and public speaking
- Poor circulation
- Feet different size
- Joint replacement
- Toenail fungus
- Bruises easily
- Low sperm count
- Male hair loss
- Having a chin dimple (?!? I kid you not. I don’t even know where to begin)
Now, we should point out here that it is easy to get excited about these sorts of association and think to oneself “Oh gosh, I have a chin dimple! I’m going to get Parkinson’s! Run for the hills!” But the thing to remember is that these are simply associations, they are not definitive causal phenotypes. Just because a person has a chin dimple (seriously?!?), it does not necessarily mean that they are going to develop Parkinson’s. Likewise, I would like to insist that readers don’t rush out and get a nose job or liposuction in a desperate attempt to ward off Parkinson’s.
Always remember that the average person on this planet has less than two arms – you only need a one or two individuals to differ from the normal for the needle to move.
And there is another note of caution regarding this new data: the researchers admit that there are certain limitations in their study, for example they can not confirm Parkinson’s diagnosis for everyone that took part (some individuals may be mis-diagnosed or not have PD at all). In addition, the analysis was largely restricted to individuals of European ancestry which raises questions of how the data may relate to folks with Parkinson’s from Asian backgrounds (for example). Finally, they report that the survey completion rates were variable across phenotypes, so some of the phenotypes may have less statistical power than others.
Still, interesting data!
What does it all mean?
Epidemiological studies – tracking and analysing large groups of people over time – are very useful tools for the medical community. They have certainly pointed out a lot of interesting aspects of Parkinson’s that were not immediately apparent (such as REM sleep behaviour disorder). And these associations that have been able to tell us something about the biology of the condition (for example, REM sleep behaviour disorder is believed to result from the loss of cells in the brain stem that inhibit the body from acting out one’s dreams while sleeping).
But there is an important feature of these sorts of studies that must be considered when looking at the results: what is the data not telling us?
The smoking and Parkinson’s association is a good example to explain what I mean here. There has always something about this association that has bugged people ever since it was first suggested: Is it smoking (or an ingredient in the tobacco) per se that infers resistance to Parkinson’s? Or is it that people who are predisposed to smoking are less likely to develop Parkinson’s?
Just as epidemiological studies are designed to consider all of the variables, we must equally consider all of the possible explanations when reviewing the data. While I’m still not sure what variables there are to consider regarding chin dimples (did the researchers insert that one just to see if anyone was actually reading their research? Is this some kind of a test?), there are some phenotypes (for example, Adventurous and Narcissism) that could explain other associations with Parkinson’s (like that whole cosmetic surgery/liposuction thing).
While some of these new phenotypes may ultimately be causalities of statistical analysis, others could tell us something new about the condition. We just need to have them independently replicated before we start getting too excited about them. We also need to see what this manuscript will look like after it has passed through the rigorous peer-review process – I am reluctant to share research here until a research report has survived that ordeal, but in this particular case I thought the findings would be of interest to the readers. I will amend this post if anything changes in the manuscript once it is actually published.
And I am open to ANY ideas or personal theories regarding that chin dimple association. Please leave them in the comments below.
The banner for today’s post was sourced from 23andMe