EDUCACIÓN PRIMARIA.

Progressive supranuclear palsy (PSP) and multiple system atrophy (MSA) are two distinct and well described neurodegenerative disorders but little is known about their prevalence and incidence in the general population. They are frequently misdiagnosed, particularly as Parkinson’s disease (PD). The diagnosis may be particularly difficult in the early stages of the disease before atypical features have clearly emerged, so that ideally patients should be regularly reassessed by movement disorder specialists (Rajput, 1991). In the United Kingdom Parkinson's Disease Society Brain Bank, neuropathological examination of patients who died with a diagnosis o f PD showed that around 6% and 5% of them actually had PSP and MSA respectively (Hughes,

1992), and among a series of 35 pathologically proven cases of MSA, 55% of the 22 cases referred from other institutions died with an erroneous diagnosis of PD

(Wenning, 1995). Over the 5 year period from 1989 to 1994, 7.9% and 9.9% of 303 brains donated had PSP and MSA, respectively (Dr. Susan Daniel, personal

communication). However, PSP and MSA patients are likely to be over-represented in studies of PD brain banks (Maraganore, 1998), and therefore extrapolating from these results to estimate their prevalence may produce very misleading estimates. Moreover, because survival in PSP and MSA is markedly shorter than in PD, such figures are a better indicator of incidence than prevalence.

There are no large population-based studies on the prevalence of PSP and MSA.

Prevalence studies of PD usually either exclude atypical parkinsonian patients or do not attempt to diagnose such patients but simply include them under the diagnosis of parkinsonism. (Tandberg, 1995; Morgante, 1992). This population-based study was primarily carried out to determine the prevalence rates of PSP and MSA.

4.2. Methods

The methods of case ascertainment and diagnosis were described in chapter 3. In brief, the computerised records o f 15 general practices (12 participating in the linkage

scheme and 3 additional practices; 14 in London and 1 in Kent) were screened for patients with the following entries: ‘Parkinson’s disease’, ‘parkinsonism’, ‘progressive supranuclear palsy’, ‘striatonigral degeneration’, ‘Shy-Drager syndrome’,

‘parkinsonism with orthostatic hypotension’, and ‘other extrapyramidal disorders, not otherwise specified’, ‘tremor’ with onset above the age of 50 years, as well as for those patients who had ever been prescribed an antiparkinsonian drug. This over-inclusive multiple approach was taken to ensure that the screening criteria had a high sensitivity (include all PSP and MSA cases) at the expense of low specificity (include lots of other conditions). Patients with only tremor beginning before the age of 50 were excluded because the risk of missing PSP or MSA patients presenting at a young age with isolated tremor without other neurological signs was considered minimal. I reviewed the computerised and hard copy records of all eligible patients. Patients were excluded from further evaluation in this part of the study if they (a) had used antiparkinsonian drugs for other indications (e.g. bromocriptine for pituitary adenoma or anticholinergic drugs taken ‘prophytactically’ in association with dopamine antagonists in patients with psychotic disorders), (b) had a known alternative cause of tremor (e.g.

hyperthyroidism), (c) had onset of parkinsonian symptoms within 6 months of at least 6 months treatment with dopamine receptor blocking drugs, (d) developed dementia before the onset of their parkinsonism, and (e) had been miscoded. All the remaining patients were contacted through their general practitioners (in all but one practice) and invited to be seen at the NHNN, the general practice, or at home. Those who did not reply were sent a reminder letter and, if they failed to respond, received a reminder telephone call.

4.2.1. Diagnosis. All patients who agreed to participate had a general and neurological interview and examination. I also administered a questionnaire specifically designed to detect signs and symptoms of PSP, MSA, or other atypical parkinsonian disorders. To enable independent diagnostic validation, a video recording of the neurological signs was made if the patient consented. The diagnosis was made according to published criteria (see chapter 3 and below) after review and discussion of each subject with NPQ.

As the diagnosis in some cases may only become evident after a period of time as new clinical features develop, patients were kept under review using three methods. All patients with a diagnosis of parkinsonism, PSP or MSA received a three monthly questionnaire on the development of atypical features and symptoms o f progression (e.g. development o f falls). Secondly, the general practitioners were asked at the end of the study whether they had noticed any new atypical features in any of their eligible patients. Finally, patients with either no definitive diagnosis or with atypical features at the first visit, or who developed them during follow-up, were reviewed and re­

examined after at least one year. As much information as possible was sought from the general practitioner and other hospitals they had attended on those patients who

declined to participate. The same diagnostic criteria were applied to these patients. 4.2.2. Diagnostic criteria. PSP (probable and possible) was diagnosed according to criteria proposed by the National Institute of Neurological Disorders and Stroke

(NINDS) and the Society for PSP, Inc. (SPSP; Litvan, 1996a), and MSA (probable and possible) according to the criteria of Quinn (Quinn, 1989; table 3 and 4).

4.2.3. Data analysis. The numbers of patients registered with each general practice by sex and age were obtained from the practice or the responsible Family Health Services Authority. Prevalence day was 1st July 1997. Crude prevalence and age-adjusted rates were calculated by direct standardisation to the 1996 European Population (Eurostat Data Shop). Crude incidence rates were calculated indirectly by dividing the prevalence rates by mean disease duration from previous publications (see below), making the assumptions o f a steady state population. 95% confidence intervals (95% Cl) were calculated for the crude rates using the Poisson formula and for the standardised rates according to the method of Smith (Smith, 1987).

4.3. Results

The total population was 121,608 people, of whom 23,859 were older than 55, and 14,272 older than 65 years. The screening process identified 679 patients, but 438 patients were excluded after reviewing their records, leaving 241 eligible patients. The reasons for exclusion were a) the use of antiparkinsonian drugs for other indications in 334 patients (76%; 286 of whom had been receiving an anticholinergic in association

with a dopamine antagonist for a psychiatric disorder), b) a known alternative cause of tremor in 51 patients (12%), c) drug-induced parkinsonism in 43 patients (10%), d) dementia before the onset o f parkinsonian symptoms in 5 patients (1%), and e) miscoding in 5 patients (1%). 202 of these patients agreed to be examined (84% participation rate), with 33 patients declining to participate, and 6 patients died before they could be seen. There were no statistically significant differences for sex (59% vs. 51% female, Chi^ value = 0.63, p=0.43) comparing patients who were not with those who were seen, but patients who were not seen were significantly older than those who were seen (79.5 years vs. 73.0 years, t-test p<0.001).

170 patients fulfilled the criteria for parkinsonism (137 patients seen, 33 not seen). Five patients fulfilled criteria for probable PSP, and one patient was classified as having possible PSP. Two patients had probable MSA, and one patient qualified for possible MSA. One further patient, who had declined to be seen, was under the care of a movement disorder specialist with a suspected diagnosis of MSA, and fulfilled criteria for possible MSA. Apart from this case, none of the other patients had been previously diagnosed with either of these two diseases (table 6), the previous diagnoses being PD (n=7), small vessel disease (n=l) and no diagnosis (n=l). Only two of the MSA patients and three o f the PSP patients had previously seen a neurologist. In addition to these ten patients, another four with parkinsonism and features atypical for idiopathic PD were seen, who did not fulfil criteria for PSP, MSA, or another specific

parkinsonian disorder (see below). All but one patient with probable PSP, who was of Asian origin, were Caucasian. There were no cases o f corticobasal degeneration or other, rare neurodegenerative disorders.

4.3.1. Prevalence. The crude and age-adjusted prevalence for PSP, if all probable and possible cases were included, was 4.9 per 100,000 (95% confidence interval 1.8 to 10.7 per 100,000) and 6.4 per 100,000 (95% confidence interval 2.3 to 10.6 per 100,000), respectively (table 7). If only those with probable PSP were included, the crude and age-adjusted prevalence of PSP was 4.1 and 5.5 per 100,000. If it was assumed that all atypical, not classifiable cases had PSP, the prevalence for PSP would be 8.2 and 11.1 per 100,000, respectively. The crude prevalence for MSA (all probable and possible

cases), was 3.3 per 100,000 (95% confidence interval 0.9 to 8.4 per 100,000) and adjusted rate was 4.4 per 100,000 (95% confidence interval 1.2 to 7.6 per 100,000) (table 7). If only the probable cases were included, the crude and age-adjusted

prevalence o f MSA were 1.6 and 2.2 per 100,000, respectively. If it was assumed that all atypical, not classifiable cases had MSA, the crude and age-adjusted prevalence for MSA would be 6.6 and 9.0, respectively.

4.3.2. Gender differences. 116 men (48.1%) and 125 women (51.9%) were eligible for the study. The rates for men and women with probable or possible PSP were almost identical - 5.0 per 100,000 and 4.8 per 100,000 based on 3 male and 3 female cases (relative rate 1.05, 95 confidence interval 0.14 to 7.8). For probable and possible MSA, the rates were 1.7 per 100,000 from men and 4.8 per 100,000 for women (relative rate 0.35, 95% confidence interval 0.01-4.3) Among the four patients with atypical, not otherwise classified parkinsonism, one was male.

4.3.3. Incidence. Using survival times o f 5.6 years for PSP derived from Litvan et al (Litvan, 1996b) and 5.3 years derived from Bower et al (Bower, 1997), indirect age- adjusted incidence rates were calculated for (probable and possible) PSP o f 1.14 and 1.21 per 100,000 per year, respectively. For probable and possible MSA, using survival times of 9.5 derived from Wenning et al (Wenning, 1994a) and 8.5 derived from Bower et al (Bower, 1997), age-adjusted incidence rates were obtained of 0.46 and 0.52 per 100,000 per year, respectively.