News from the Land of Milk and Honey

I had the good fortune to attend a conference on Precision Medicine and Ageing today. This brought some of the leading scientists from the UK and Israel together, as part of the British Council’s Britain Israel Science and Innovation Network. The day included speakers talking about the latest breakthroughs in genomics, neuroscience, cellular biology and many other fields of scientific endeavour and how they relate to the biology of ageing (in health as well as disease) and how these breakthroughs are bringing us closer to the ‘promised land’ of Personalised, or Precision medicine.

The talk that I found the most interesting came from Dr Maya Leventer-Roberts, from the Clalit Research Institute. This is not a university department, but the in-house research arm of the largest of Israels four health providers. They have an unparallelled treasure-trove of health data. She explained that all Israeli citizens belong to one of these four organisations who provide all aspects of care: GP, pharmacy, testing laboratories, hospitals – you name it. What’s more, very few people change providers, so they have a well-coded database of people’s wellness, illness and health behaviour at all stages of life.

I have since found a paper from last year which uses some of their data to examine a relatively neglected aspect of Parkinson’s – depression. Unfortunatley a huge proportion of people with Parkinson’s struggle with depression. We also know that late onset depression is a risk factor for Parkinson’s. 

In this study, they looked at mortality rates in Parkinson’s and the association with how adherent people were with antidepressents. They analysed the data of 10,000 people with a diagnosis of Parkinson’s as well as at least one prescription of an antidepressant. They then categorised people into how frequently they tookthe antidepressants by comparing their prescriptions from the doctor with the number of purchases of antidepressants from the pharmacies. Finally they looked at death by any cause over the 4 year study period.

They found that the people who had Parkinson’s and depression who took their antidepressants the least were 43% more likely to die within the 4 year period, compared to those who took it the most regularly. The increased mortality was exacerbated by having an increasing number of other medical problems.

I must stress the old adage: association does not equal causation. This kind of retrospective observational study cannot provethat one thing causes another. However, it is an important reminder that depression is an important aspect of Parkinson’s, and that poorly managed depression is associated with increased mortality. It also reminds me that older people, with Parkinson’s, depression and perhaps four or five other medical conditions may be on 15-20 or more tablets a day. What can I do as a clinician to help them remember what each is for, and why each is important to take? What can we do as a society to help people with depression with both drug and non-drug treatment options? How can we better prepare ourselves, individually and societally, for ageing and the only true inevitability of life?

RNR

Shoval G, Stubbs B, Balicer RD, Feldman B, Hoshen M, Zalsman G, et al. 

Low adherence to antidepressants is associated with increased mortality in Parkinson disease patients. 

Parkinsonism Relat Disord. 2017 Oct;43:92–6.

INTRODUCTION:The purpose of this study was to evaluate the relationship between adherence to antidepressants (AD) and all-cause mortality in a population-based cohort of patients with Parkinson's Disease (PD).

METHODS:From a database of more than 4 million people, 8553 patients with PD who purchased an AD at least once between the years 2008-2011 were retrospectively followed for all-cause mortality over 4-years. Adherence was measured as a ratio between dispensed and prescribed durations and was modeled as: non-adherence (<20%, n = 1566), poor (20%-50%, n = 1184), moderate (50%-80%, n = 1584), and good (>80%, n = 4219) adherence. Multivariable survival analyses adjusted for demographic and clinical variables including physical comorbidities known to influence mortality were conducted, however there was no adjustment for other psychiatric disorders and medications.

RESULTS:Unadjusted mortality rates were 20.4%, 25.1%, 23.4% and 25.6% in those classified as non-adherent, poor, moderate and good adherence respectively (χ2 = 18.45, p < 0.0001). The non-adherent and poor adherence groups had significantly increased adjusted mortality hazard ratios (HR) of 1.43 (CI: 1.26-1.62) and 1.26 (CI: 1.1-1.44) respectively compared to the good adherence group. Using the same model, the adjusted HR for death among males was 1.49 [95% CI: 1.36-1.62] compared to females. People with PD and Charslon's Comorbidity Index score of 3-4 (HR 1.3, P < 0.001) and 5+ (HR 1.78, P < 0.001) were more likely to die than those with 0-2 comorbidities.

CONCLUSIONS:Our findings suggest that poor adherence to AD is associated with increased all-cause mortality in people with PD. Given the high prevalence of depression and AD effectiveness, efforts to promote adherence should be prioritized in clinical practice.

Application of the Movement Disorder Society prodromal criteria in healthy G2019S-LRRK2 carriers

This is an interesting application of the MDS Task Force criteria for prodromal PD. Here the Ashkenazi Jewish consortium applied the MDS criteria to non-manifesting carriers and non-carriers of the G2019S-LRRK2 mutation. 

Despite relatively small numbers of participants being diagnosed with PD during the follow-up period, they report impressive specificity of the criteria for idenitfying those in the pre-diagnostic (prodromal) phase of PD. 8 of the 10 patients (80%) that were diagnosed during follow-up were in the highest probability group at baseline. Unfortunately I could not find the raw data in the manuscript to determine how they arrived at 91.8 % specificity but working backwards from the reported specificity and positive predictive value I think there must have 9 out of 110 (8.2%) who were in the high probability group  at baseline, but did not subsequently get diagnosed with PD during follow-up. It will be interesting to see what happens to those participants with more time.

The positive predictive value (PPV) tells us about the probability of getting a diagnosis of PD over time if you are in the highest risk group at baseline. Based on the follow-up so far, 47% of people that screened positive at baseline have been diagnosed with PD. Whilst that's approximately 50:50 - if you imagine a scenario where a preventive treatment were available and the toxicity or side effects of that treatment were low then perhaps 50% PPV for a screening test is not widely inappropriate. Of course these data only apply to carriers of the G2019S LRRK2 mutation which in itself puts people at significantly higher risk of PD. The MDS Task Force criteria have been applied to other at risk cohorts, but we are in the final stages of applying them to the PREDICT-PD population-based cohort... watch this space.

- Alastair Noyce

Mov Disord. 2018 Mar 30. doi: 10.1002/mds.27342. [Epub ahead of print]
Mirelman A, Saunders-Pullman R, Alcalay RN, Shustak S, Thaler A, Gurevich T, Raymond D, Mejia-Santana H, Orbe Reilly M, Ozelius L, Clark L, Gana-Weisz M, Bar-Shira A, Orr-Utreger A, Bressman SB, Marder K, Giladi N; AJ LRRK2 Consortium.

https://onlinelibrary.wiley.com/doi/full/10.1002/mds.27342

BACKGROUND: In 2015, the International Parkinson and Movement Disorder Society Task Force recommended research criteria for the estimation of prodromal PD.

OBJECTIVES: We aimed to evaluate, for the first time, the criteria in first-degree relatives of Ashkenazi Jewish G2019S-LRRK2 PD patients, who are considered a population at risk for developing PD, and assess the sensitivity and specificity of the criteria in identifying phenoconverters.

METHODS: Participants were evaluated longitudinally over a period of 5 years (average follow-up: 49.2 ± 12.3 months). Likelihood ratios and probability estimations were calculated based on the International Parkinson and Movement Disorder Society Research Criteria for Prodromal Parkinson's Disease markers and examined for each assessment point.

RESULTS: One hundred twenty healthy carriers (49.53 ± 13.4 years; 54% female) and 111 healthy noncarriers (48.43 ± 15.79 years; 49% female) participated in this study. Probability scores were significantly higher in healthy carriers than healthy noncarriers (P < 0.0001). Of the 20 participants (8.6%) who met criteria for probable prodromal PD at baseline, 17 were healthy carriers. Participants who reached the threshold were older (P < 0.0001), had higher UPDRS-III (P < 0.001), lower cognitive function (P = 0.001), and more nonmotor symptoms (P < 0.0001), compared to those who did not. Ten participants were diagnosed with incident PD within 5 years from baseline resulting in a specificity of 91.82% (95% confidence interval: 86.69-96.94), sensitivity of 80% (95% confidence interval: 55.21-100), positive predictive value of 47.06% (95% confidence interval: 23.33-70.79), and negative predictive value of 98.06% (95% confidence interval: 95.39-100). All 10 phenoconvertors were G2019S-LRRK2 carriers.

CONCLUSIONS: The results showed the utility of using the criteria and high sensitivity and specificity in identifying prodromal PD in this high-risk unique cohort. These results may be valuable for future disease modification clinical trials. © 2018 International Parkinson and Movement Disorder Society.

 © 2018 International Parkinson and Movement Disorder Society.

Detecting anxiety in individuals with Parkinson disease: A systematic review

This is a new systematic review of studies assessing the diagnostic accuracy of anxiety scales in people with PD. It is important because anxiety is a common feature of Parkinson's both before and after the diagnosis. The authors found that overall 30% of people with Parkinson's meet the criteria for anxiety, but I have to say that in my experience I would put it closer to 50%. Maybe this reflects a bias in the patients I see??

The properties of the scales available for assessing anxiety are important. At the beginning of PREDICT-PD, we went through a pain-staking process of selection and in the end arrived at the Hospital Anxiety Depression Scale (HADS). In fact, at that time we knew that it was not the very best Anxiety scale or the very best Depression scale in terms of diagnostic accuracy, but it did combine assessment of both problems in one validated questionnaire, it was feasible to self-administer and it was relatively short. Nearly seven years on I am still happy with that decision but we may now test other scales within the PREDICT-PD programme too. Depression and anxiety are two features that are likely to pre-date a diagnosis of PD by a long time and it is critically important to be using the best tools to identify them both. They also have a dramatic impact on quality of life and are treatable, which are two good reasons to be trying to pick them up in people that don't have a diagnosis because there are likely to be significant health benefits...

Alastair Noyce

Neurology. 2017 Dec 6. pii: 10.1212/WNL.0000000000004771. doi: 10.1212/WNL.0000000000004771. [Epub ahead of print]
Mele B, Holroyd-Leduc J, Smith EE, Pringsheim T, Ismail Z, Goodarzi Z.

http://n.neurology.org/content/early/2017/12/06/WNL.0000000000004771.long

OBJECTIVE: To examine diagnostic accuracy of anxiety detection tools compared with a gold standard in outpatient settings among adults with Parkinson disease (PD).

METHODS: A systematic review was conducted. MEDLINE, EMABASE, PsycINFO, and Cochrane Database of Systematic Reviews were searched to April 7, 2017. Prevalence of anxiety and diagnostic accuracy measures including sensitivity, specificity, and likelihood ratios were gathered. Pooled prevalence of anxiety was calculated using Mantel-Haenszel-weighted DerSimonian and Laird models.

RESULTS: A total of 6,300 citations were reviewed with 6 full-text articles included for synthesis. Tools included within this study were the Beck Anxiety Inventory, Geriatric Anxiety Inventory (GAI), Hamilton Anxiety Rating Scale, Hospital Anxiety and Depression Scale-Anxiety, Parkinson's Anxiety Scale (PAS), and Mini-Social Phobia Inventory. Anxiety diagnoses made included generalized anxiety disorder, social phobia, and any anxiety type. Pooled prevalence of anxiety was 30.1% (95% confidence interval 26.1%-34.0%). The GAI had the best-reported sensitivity of 0.86 and specificity of 0.88. The observer-rated PAS had a sensitivity of 0.71 and the highest specificity of 0.91.

CONCLUSIONS: While there are 6 tools validated for anxiety screening in PD populations, most tools are only validated in single studies. The GAI is brief and easy to use, with a good balance of sensitivity and specificity. The PAS was specifically developed for PD, is brief, and has self-/observer-rated scales, but with lower sensitivity. Health care practitioners involved in PD care need to be aware of available validated tools and choose one that fits their practice.