Epi in the News

First time trying one of these - hope i don't fall flat on my face

1. Cohort study
2. Swedish citizens/residents registered with the National Health Registry who do NOT suffer from Parkinson's disease.
3. The article worded it pretty poorly - but I'm guessing 3 exposure groups ? Vitamin C only or E only , Vitamin C AND E, and those with Beta-carotene, those with NEACH - poorly defined in the article - not sure about the actual study or how they presented it

4. The participants in the highest third consumption level of vitamin C AND E had an 84.2% reduce risk compared to the highest third consumption of only C or E? or Those who only consume Vitamin C or E only are at an increased 18% risk of developing Parkinson's compared to those who consume both.
5. speculative reasons: no dose response seen until that upper on third
no strong evidence (high p value, 1 inclusive CI?) to support until stratifying at different levels of consumption.
6. Maybe? Can't do a trial - technically unethical - potentially an animal study - too complex of a disease to create conditions leading to its emergence - potentially a case control study to retrospectively establish that relationship to form a link?
7. 5.57 cases of parkinson's per 10,000 person-years

Cheers!

Hi everybody...
Thank you Mustafa for your post... it was interesting... the brainstorming in "Epi in the news" has always been extremely beneficial for me so welcome to the club...:-)

Coming to the current article:

1) The study design would be cohort study. My reasoning is that the investigators have started with ppl without the outcome of interest, followed them over a period of time and looked for the onset of the outcome of interest. The measure of association looked at is the risk of the outcome in the exposure groups (here its the lack of certain vitamins in the diet)- which is what is done in cohort studies.

2) The inclusion criteria would be ppl who have not been diagnosed with the outcome (Parkinsons disease) or who don't have known risk factors to the disease like maybe family history. They would have started off with a younger age group since their intention was to follow long term dietary habits. How plausible is it that the researchers got this aspect of recruitment right? Well, Parkinsons has a silent period and has a 10-20% early onset variant that could be missed. So some participants might already be having the pathology but just not diagnosed when they entered the study, and its well known that the disease is often quite hard to diagnose when symptoms are hidden or manifest only during certain acts. So, the apparent temporal association might be misleading.

3) They divided the exposure groups based on the amount of vitamin intake. The first group constituted 1/3rd of the total with the lowest intake of vitamins, and the last one with highest intake. So I am guessing that should make it 3 groups- lowest 1/3rd, middle 1/3rd and highest 1/3rd. These groups were defined for vitamin c intake, vitamin E intake, beta carotene and NEAC

4) "Those in the highest one-third in consumption of both vitamins together had a 38 percent reduced risk."
Here the RR would be 0.62 (1- 0.38= 0.62).
Incidentally this is the RR for combined Vitamin C and E which has not been reported in the abstract of the original paper- so the article here has chosen to report that figure.. wonder why?

5) I think maybe the RR was not significant when comparing the middle 1/3 with the higher or lower 1/3. The article of course is pushing to make the claim that supplementing the diet with vitamin C and E can potentially prevent Parkinson's- so they have highlighted only the RRs that were significant..

6) Well, the numbers are really good- 41000 + ppl in the study. It is definitely a start. I would like to actually read the whole article- because my biggest concern is how did they get this information about intake of vitamin E and C (and others) and how reliable is the classification of the exposure. We definitely need a trial to make any definitive conclusion, where the exposure can be controlled and properly classified. I dont agree with Mustafa here coz an equipoise still exists- an association between dietary content and Parkinson's has not been established so a trial can be planned- however the major problem here would be long latent and hidden period of the disease hence making it an unreasonably long and expensive trial.
Having said that, I like the last sentence in the article. Vitamins are anyways essential in our diet and they have been conclusively shown to have a preventive role in many other acute diseases. So, regardless of what it does to onset of Parkinson's, maybe ppl should be advised to make sure to have a well balanced diet inclusive of these vitamins. The question is will this information really change what is already being advised? I doubt it..

7) There were 465 cases among 41058 participants of the study.
The incidence risk is 465/41058= 1.1%
so in simple words:

" 1.1 percent of the participants who entered the study developed Parkinson's disease during the study time period of 18 years "
I am not sure if rate can be used here because we don't have the entire population figure- again this is one of my never ending confusions- still not able to decide whether to use risk or rate..

Fathima

1. Cohort study
2.People would need to be good at recording their diet accurately and not to have Parkinson's disease when they entered the study. Looking at the original paper it seems that participants diet was only assessed once at baseline. Participants are therefore assumed not to have changed their diet over the 18 years of the study. This does not strike me as a plausible assumption.
3. 4 exposure groups - vitamin C, E, beta-carotene, NEAC
4. The risk ratio will be [(Number of highest third vitamin intake people developing PD/total number of highest third vitamin intake people)/(Number of lowest third vitamin intake people devloping PD/total number of lowest third vitamin intake people)] x 100%
5. We imagine that the middle one third had an intermediate risk. If they had been reported and if the top third of vitamin intake people were compared to the bottom two thirds the relative risk would have been less dramatic and this would have made a less interesting newspaper article.
6. We do need a trial. Although the researchers have tried to adjust for relevant factors it is never possible to do this fully - residual confounding is likely. One of the many questions arising from this study is whether it is the vitamin C and E themselves that are associated with a lower risk of developing PD or whether having a healthy diet (and all its associations such as perhaps being wealthier, more health conscious etc etc) is the relevant factor. This could be achieved by giving people vitamin supplements or placebo in a randomised controlled trial.
7. Incidence = (number developing PD/number in the cohort)/years of study = (465/41058)/18 = 62.9/100 000/year. Given that Parkinson's disease is an age-related condition this is rather a meaningless rate if we do not say how old the cohort was when they entered the study.

Judith