Past exposure to neuroleptics is associated with the risk of developing Parkinson's disease
Question: What is the long-term risk of developing Parkinson's disease after past exposure to neuroleptics?
People: A total of 3726 people aged 65 years and over (41.8% male, mean age 74.9 years) who were living at home and did not have parkinsonism at baseline. Participants were randomly selected from the electoral rolls with stratification by age, gender and size of the demographic unit.
Setting: Two districts, South West France; 1988–2004.
Risk factors: Past use of neuroleptics (dopamine receptor antagonists, such as flunarizine or cinnarizine). Drugs used regularly or daily were assessed by questionnaire at baseline, and 3, 5, 8, 10, 13 and 15-years’ follow-up and also by visual inspection of participants’ medications.
Outcomes: Probable Parkinson's disease (PD, diagnosed according to the UK PD Society Brain Bank (UKPDSBB) criteria) and parkinsonism and other parkinsonian syndromes (such as vascular parkinsonism and drug-induced parkinsonism (DIP)). DIP was defined as parkinsonism and current or previous use of dopamine receptor antagonists in the past 3 months to a year (depending on the drug used) which remitted on drug withdrawal. Persistent DIP was excluded (parkinsonism and causative drug use at all study visits). Incident cases of parkinsonism were ascertained by a two-step process: first by screening based on either spontaneous diagnosis declaration or use of antiparkinsonian drugs, complaints of slowness or tremor, or positive screening for cardinal signs; and second, by analysing complementary data from the patients’ general practitioners or specialists. Each case file was reviewed for diagnosis classification by a diagnosis consensus board composed of three senior neurologists.
Over the 15 years, 2991 participants (79.2%) were followed up at least once, 735 participants (19.5%) died before the first follow-up. During follow-up 144 incident cases of parkinsonism were identified, of which 27 (18.8%) were persistent drug-induced parkinsonism and excluded from the analysis. Among the remaining 117 cases of incident parkinsonism, 43 had probable PD. The remaining 74 cases were: 29 of dementia with Lewy bodies, 19 of vascular dementia, 2 of parkinsonism complicating Alzheimer's disease and 24 of parkinsonism that fulfilled UKPDSBB criteria but had an unknown l-dopa response. Among the participants 516 individuals (17.3%) had been exposed to neuroleptics without presenting with persistent DIP. Before the diagnosis, the frequency of exposure to neuroleptics was 22.2% among individuals with parkinsonism, 32.6% in probable PD cases and 16.6% among those without parkinsonism. After adjusting for gender and occupation, past exposure to neuroleptics was associated with an increased risk of parkinsonism and probable PD (parkinsonism: RR 1.65, 95% CI 1.05 to 2.58, p=0.03; probable PD: RR 3.16, 95% CI 1.65 to 6.04, p=0.0005). When split by type of drug, benzamides and phenothiazines were both associated with an increased risk of probable PD (benzamides: RR 3.65, 95% CI 1.41 to 9.45, p=0.008; phenothiazines: RR 2.59, 95% CI 1.23 to 5.43, p=0.01). The association between the individual drug types and parkinsonism did not reach statistical significance (benzamides: RR 2.05, 95% CI 0.99 to 4.24, p=0.05; phenothiazines: RR 1.48, 95% CI 0.88 to 2.49, p=0.14).
Past exposure to neuroleptics among older adults is significantly associated with the risk of incident Parkinson's disease.
Sources of funding ISPEN France, NOVARTIS Pharma France, and the Caisse Nationale de Solidarité et d'Autonomie.
Large numbers of risk factors have been reported for idiopathic Parkinson's disease (PD). Most have been identified in retrospective case–control studies. Some have emerged from large-scale prospective studies aimed at multiple health outcomes. Dopamine receptor blocking drugs (DRBD) have not raised a concern until now.
Long-term use of DRBD has not been found to cause neuronal damage. The dopaminergic system in the brain is, however, subject to long-term, probably permanent changes induced by exposure to DRBD, as evidenced by tardive dyskinesia. The hypothesis that neuroleptics contribute to PD risk rather than accelerated PD identification, therefore, has a plausible foundation. There is, however, long-term experience with DRBD use in schizophrenics and reports on their developing PD are rare.1 While this certainly may be due to a number of biases and confounders, the lack of a recognised problem provides a skeptical backdrop to this study.
This prospective study was ambitious, involving patient evaluations over 15 years by trained staff. No large, complicated, multiyear undertaking such as this could be perfect. The protocol is subject to different critiques by different critics. The first problem represents the sample, which was drawn from electoral rolls. The second is the recruitment rate of 60%, of whom 41% were men. More importantly, parkinsonism screening was made initially via questionnaire, buttressed by trained personnel who were psychologists, not neurologists. The diagnosis of PD was made by movement specialists based on chart review, not clinical examination. The positive association found at 3 months off DRBD became non-significant when the withdrawal time between drug use and identification of PD was set at 2 years, possibly due to small numbers. Although drug-induced parkinsonism (DIP) usually resolves within a few months of drug discontinuation, it may last many years,2 and after several years, some progression of parkinsonism may occur due to aging itself,3 confounding a diagnosis of PD.
While there are many reasons for neuroleptics to be avoided in the elderly, the prospect of their increasing the risk of developing idiopathic PD is not clearly one of them.
Competing interests None.