Accuracy Of Clinical Diagnosis Of Progressive Supranuclear Palsy:A Systematic Review And Meta-analysis

Background:Progressive supranuclear palsy?PSP?is a neuropathologically defined disease presenting with a broad spectrum of clinical phenotypes.We aimed to search for reliable and meaningful diagnostic tools to improve the diagnostic accuracy of clinical diagnosis of PSP by a systematic review and Meta-analysis.Methods:We searched for articles published between January 1996 and May 2017.We included studies reporting diagnostic parameters regarding clinical diagnosis of PSP with crude data.In the first part,the selected studies were recruited to make recommendations about how to improve the accuracy of clinical diagnosis of PSP.In the second part,we focused on how to distinguish between PSP and other parkinsonian disorders through cognitive&neuropsychology test,clinical features,cerebrospinal fluid?CSF?proteins and imageological examination.Results:We selected 74 studies,including 19 studies using different diagnostic criteria to evaluate the diagnostic accuracy of available clinical diagnosis of PSP in the first part.The pooled accuracy of 19 studies was 0.876?95%confidence interval?CI?:0.856,0.896?;In the subgroups analysis based on the different diagnostic criteria,accuracy was 0.886?95%CI:0.863,0.908?for NINDS-SPSP Consensus criteria 1996,0.863?95%CI:0.791,0.935?for Litvan et al.Optimal criteria,and 0.843?95%CI:0.795,0.891?for clinical features;In the subgroups analysis based on diagnosed as possible PSP and probable PSP,accuracy was 0.856?95%CI:0.822,0.890?for probable PSP and 0.936?95%CI:0.898,0.974?for possible PSP.In the subgroups analysis based on the differences in type of gold standard,accuracy was 0.912?95%CI:0.878,0.945?for pathology,0.866?95%CI:0.836,0.895?for imaging,and 0.888?95%CI:0.859,0.918?for CSF examine.The subgroup analysis based on diagnose by experts have higher pooled accuracy compared to those diagnose by non-experts?0.933,95%CI:0.914,0.951 vs 0.840,95%CI:0.813,0.867?.Moreover,PSP diagnosis in the late phase of disease may improve the accuracy compared to PSP diagnosis in the early phase of disease?0747,95%CI:0.572,0.923 vs 0.704,95%CI:0.573,0.835?.In the second part,the core biomarkers differentiated PSP from PD with good performance:the Unified Parkinson's Disease Rating Scale?UDPRS?II?activities of daily life?scores?5.66,95%CI:2.99,8.33?,UDPRS III?motor examination?scores?7.50,95%CI:6.21,8.80?,Hoehn and Yahr?HY?scale?0.85,95%CI:0.61,1.09?,Mini-Mental State Examination?MMSE?score?-2.81,95%CI:-4.37,-1.24?,Frontal Assessment Battery?FAB?test?-6.24,95%CI:-9.89,-2.60?;gastrointestinal symptoms?3.11,95%CI:1.43,6.75?,apathy?4.57,95%CI:2.32,9.01?,falls?36.57 95%CI:5.26,254.42?,gaze palsy?179.00,95%CI:15.01,2134.50?,blepharospasm?3.33,95%CI:1.29,8.59?,tremor?0.05,95%CI:0.01,0.45?,orthostatic hypertension?0.23,95%CI:0.09,0.57?,symptomatic orthostatic hypertension?0.15,95%CI:0.04,0.62?;total-tau?-34.84,95%CI:-68.30,-1.39?,p-tau?-3.12,95%CI:-4.82,-1.43?,neurofilament light chain?NFL??1723.51,95%CI:785.79,2661.22?,?-synuclein?0.15,95%CI:0.03,0.27?;putamen?-1.75,95%CI:-3.20,-0.30?,caudate?-2.31,95%CI:-3.62,-1.00?,midbrain volume?-1.25,95%CI:-1.61,-0.90?,third ventricle?0.83,95%CI:0.56,1.11?.Differentiation between PSP and MSA was also strong?OR-5.46 for UDPRS II?activities of daily life?scores,-5.75 for UDPRS total scores,-1.71 for MMSE score,9.99for falls,0.07 for urinary symptoms,0.09 for orthostatic hypertension,0.23 for symptomatic orthostatic hypertension,-821.32 for NFL,0.25 for?-synuclei,-1.61 for putamen,-1.70 for caudate,and 0.60 for third ventricle?.Furthermore,falls?47.03,95%CI:11.99,184.51?,A?₄₂?132.40,95%CI:82.62,182.19?,and p-tau?-11.09,95%CI:-18.72,-3.45?had large effect sizes when differentiating between PSP and DLB.Finally,FAB test?3.30,95%CI:2.71,3.88?,Clinical Dementia Rating?CDR?score?1.31,95%CI:0.14,2.49?,falls?8.87,95%CI:2.97,26.51?,myoclonus?0.03,95%CI:0.00,0.17?,axial rigidity?80.90,95%CI:9.02,725.78?,and total intracranial volume?2.49,95%CI:12.15,112.83?can also be used to distinguish between PSP and CBD.Conclusion:In the first part,the overall validity of clinical diagnosis of PSP is not satisfying,with high specificity but relatively low sensitivity.Imaging and CSF biomarkers are urgently needed to improve the accuracy of clinical diagnosis of PSP in vivo.In the second part,cognitive&neuropsychology test,clinical features,CSF proteins and imageological examination are useful in discriminating PSP from other parkinsonian disorders.