Prevalence of PSP and MSA

The Parkinson's-Reversing Breakthrough

Treatment Options for Parkinsons

Get Instant Access

Measures of the prevalence of PSP vary from 0.97 to 6.54 per 100,000 (eight studies) and for MSA from 2.29 to 39.3 per 100,000 (five studies) (Table 1). A study from Sicily (8) reported one of the highest rates of 28.6 per 100,000 but this was for unspecified parkinsonism, so this category may have included cases other than MSA and PSP. The results presented in Table 1 are simple crude rates for the whole population rather than age-standardized rates. Though some studies do report such standardized rates, each study generally uses a different standard population and some studies fail to present age-specific rates making it impossible to restandardize the rates to a single population. The purpose of standardization is to remove any confounding effect owing to the age structure of the population. As all the studies except one from Libya (7) have been undertaken in a developed world population, it is likely that such confounding is not too large. In fact, differences between the crude and standardized rates are often small. For example, the crude rate for the New Jersey study is 1.38 per 100,000 and after standardization this increases to 1.39 per 100,000 (6). In the London study, the crude PSP rate of 4.9 per 100,000 increased to 6.4 per 100,000 (14). However, the incidence rate from Benghazi, Libya, is misleadingly small as a large proportion of this population will be under 55 yr of age and standardizing this rate to a European population will certainly increase this rate by a large degree.

Despite the various different diagnostic criteria, the rates for PSP and MSA across all studies are not too dissimilar. One obvious observation is that studies with very large populations, e.g., New Jersey (6) or United Kingdom (15), produce lower prevalence estimates whereas very small populations produce high rates. The effect of varying population size and hence intensity of case finding is most elegantly demonstrated by the "Russian Doll Method" method employed by Nath and colleagues (15). This general pattern is also well noted with prevalence studies for multiple sclerosis

(16). This is unsurprising because, although large populations give precise estimates, they are likely to be biased downward as it is easier to miss a proportion of cases. On the other hand, smaller populations will enable much more thorough case ascertainment and hence less biased but imprecise estimates. The highest prevalence rate (39.3 per 100,000) was observed in a study from rural Bavaria (10). However being based on only three cases in a very small population, its lower 95% confidence interval is still 8.1 per 100,000, which will overlap with most of the other estimates. Almost all the studies except that from Nath and colleagues (15) estimate rates based on 11 or fewer cases, which demonstrates the problem with studying rare neurological disorders.

Some studies have specifically aimed to detect cases of PSP and/or MSA whereas others have been generic prevalence studies for parkinsonism. Interestingly, there is little difference between the estimates for these different studies. Perhaps more surprising is that studies that have screened a whole population (9,10) did not find higher rates than those using existing medical records or other databases. It is well recognized in the Parkinson's disease (PD) literature that a proportion of PD cases will be detected de novo by the screening procedure, though this varies across European centers

(17). It is possible that this "clinical iceberg" phenomenon is less important for PSP and MSA because their symptoms and disease progression make them less likely to be undetected. However, studies aimed at parkinsonism may underestimate rates of MSA since cases with predominantly cerebellar features may be undetected. Similarly, excluding patients who became demented before the onset of parkinsonism (14) will assist in the exclusion of dementia with Lewy bodies, in which a supranuclear gaze palsy may occur, but may also exclude cases of PSP.

Table 1

Prevalence Studies of PSP and MSA

Lead Author (Ref. No.)

Year of Publication

Generic (G) or Specific (S)


Population Size (Total Estimated)

Case Ascertainment

Diagnostic Criteria

PSP Prevalence per 100,000 (No. of Cases)

MSA Prevalence per 100,000 (No. of Cases)

Golbe (6)

Was this article helpful?

0 0

Post a comment