Background PSP is a neuropathologically defined disease entity. Clinical diagnostic criteria, published in 1996 by the National Institute of Neurological Disorders and Stroke/Society for PSP, have excellent specificity, but their sensitivity is limited for variant PSP syndromes with presentations other than Richardson’s syndrome. Objective We aimed to provide an evidence- and consensus-based revision of the clinical diagnostic criteria for PSP. Methods We searched the PubMed, Cochrane, Medline, and PSYCInfo databases for articles published in English since 1996, using postmortem diagnosis or highly specific clinical criteria as the diagnostic standard. Second, we generated retrospective standardized clinical data from patients with autopsy-confirmed PSP and control diseases. On this basis, diagnostic criteria were drafted, optimized in two modified Delphi evaluations, submitted to structured discussions with consensus procedures during a 2-day meeting, and refined in three further Delphi rounds. Results Defined clinical, imaging, laboratory, and genetic findings serve as mandatory basic features, mandatory exclusion criteria, or context-dependent exclusion criteria. We identified four functional domains (ocular motor dysfunction, postural instability, akinesia, and cognitive dysfunction) as clinical predictors of PSP. Within each of these domains, we propose three clinical features that contribute different levels of diagnostic certainty. Specific combinations of these features define the diagnostic criteria, stratified by three degrees of diagnostic certainty (probable PSP, possible PSP, and suggestive of PSP). Clinical clues and imaging findings represent supportive features. Conclusions Here, we present new criteria aimed to optimize early, sensitive, and specific clinical diagnosis of PSP on the basis of currently available evidence.

Progressive supranuclear palsy (PSP) is a 4R-tauopathy predominated by subcortical pathology in neurons, astrocytes, and oligodendroglia associated with various clinical phenotypes. In the present international study, we addressed the question of whether or not sequential distribution patterns can be recognized for PSP pathology. We evaluated heat maps and distribution patterns of neuronal, astroglial, and oligodendroglial tau pathologies and their combinations in different clinical subtypes of PSP in postmortem brains. We used conditional probability and logistic regression to model the sequential distribution of tau pathologies across different brain regions. Tau pathology uniformly predominates in the neurons of the pallido-nigroluysian axis in different clinical subtypes. However, clinical subtypes are distinguished not only by total tau load but rather cell-type (neuronal versus glial) specific vulnerability patterns of brain regions suggesting distinct dynamics or circuitspecific segregation of propagation of tau pathologies. For Richardson syndrome (n = 81) we recognize six sequential steps of involvement of brain regions by the combination of cellular tau pathologies. This is translated to six stages for the practical neuropathological diagnosis by the evaluation of the subthalamic nucleus, globus pallidus, striatum, cerebellum with dentate nucleus, and frontal and occipital cortices. This system can be applied to further clinical subtypes by emphasizing whether they show caudal (cerebellum/dentate nucleus) or rostral (cortical) predominant, or both types of pattern. Defining cell-specific stages of tau pathology helps to identify preclinical or early-stage cases for the better understanding of early pathogenic events, has implications for understanding the clinical subtype-specific dynamics of disease-propagation, and informs tau-neuroimaging on distribution patterns.

The phenotypic variability of progressive supranuclear palsy (PSP) may account for its frequent misdiagnosis, in particular in early stages of the disease. However, large multicenter studies to define the frequency and natural history of PSP phenotypes are missing. In a cohort of 100 autopsy-confirmed patients we studied the phenotypic spectrum of PSP by retrospective chart review. Patients were derived from five brain banks with expertise in neurodegenerative disorders with referrals from multiple academic hospitals. The clinical characteristics of the 100 cases showed remarkable heterogeneity. Most strikingly, only 24% of cases presented as Richardson's Syndrome (RS), and more than half of the cases either showed overlapping features of several predescribed phenotypes, or features not fitting proposed classification criteria for PSP phenotypes. Classification of patients according to predominant clinical features in the first 2 years of the disease course allowed a more comprehensive description of the phenotypic spectrum. These predominance types differed significantly with regard to survival time and frequency of cognitive deficits. In summary, the phenotypic spectrum of PSP may be broader and more variable than previously described in single-center studies. Thus, too strict clinical criteria defining distinct phenotypes may not reflect this variability. A more pragmatic clinical approach using predominance types could potentially be more helpful in the early recognition of and for making prognostic predictions for these patients. Given the limitations arising from the retrospective nature of this analysis, a systematic validation in a prospective cohort study is imperative.